msan621_ml1/ML-Lecture5.2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ML-Lecture5.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('/Users/tlee010/Desktop/github_repos/fastai/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import sys\n",
    "\n",
    "# or wherever you have saved the repo\n",
    "sys.path.append('/Users/tlee010/Desktop/github_repos/fastai/')\n",
    "\n",
    "from fastai.imports import *\n",
    "from fastai.structured import *\n",
    "from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier\n",
    "from IPython.display import display\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Previously: Random Forests in General\n",
    "\n",
    "Downside - can't extrapolate very well. \n",
    "\n",
    "They are a great starting point. There's a huge catalog of machine learning tools out there. Books don't curate it down. Why would you use a Support Vector Machine today? Theoretically elegant, and exciting previously, but practically dont see them having any place today.\n",
    "\n",
    "Hard to write about Random Forests, since its hard to look at it theoretically. GBM / XGBOOST etc. Now industry is more invested in the algorithms and are doing more work.\n",
    "\n",
    "Packages being used today , top 5 or 6 did not exist more than 3 years ago.\n",
    "\n",
    "**Examples**:\n",
    "- Feature interactions for RF?\n",
    "\n",
    "##### How big should your validation set be? \n",
    "\n",
    "First need to answer the question, how accurate does this algorithm need to be? \n",
    "\n",
    "|parameter | model 1| model 2|\n",
    "|-----------|--------|-------|\n",
    "|n|2000|2000|\n",
    "|acc|99.4%|99.2%|\n",
    "|wrong|12|16|\n",
    "|cats|50%|50%|\n",
    "|wrong cats|6|8|\n",
    "\n",
    "There's a practical consideration here. 0.2% could be a big deal depending on what the application. 22 is where the t distribution turns into a normal distribution.\n",
    "\n",
    "**Rule of Thumb 1** - Once we have a n > 22, it behaves more normally. So 22 out of 1000 etc. \n",
    "\n",
    "**Rule of Thumb 2** - If you have binomial data\n",
    "\n",
    "|Metric |Binomial |\n",
    "|----|----|\n",
    "|np | mean|\n",
    "|np(1-p)|std|\n",
    "|std/sqrt(n)|SE|\n",
    "\n",
    "**Rule of Thumb 3** - training set should have an equal number of the rare object, you should over sample until its the same size as the common class. There's also a way to weight the uncommon class, so its picked more often\n",
    "\n",
    "** one shot learning ** - recognize somethign that you have only seen once\n",
    "** zero shot learning ** - not covered in this class, but recognize something you never seen before"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Returning to our Random Forest Class\n",
    "\n",
    "HW - add all the other RF interpretations that we have already covered"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "import complete\n"
     ]
    }
   ],
   "source": [
    "import feather\n",
    "df_raw = feather.read_dataframe('/Users/tlee010/kaggle/bulldozers/bulldozers-raw')\n",
    "print('import complete')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df_trn, y_trn, nas = proc_df(df_raw, 'SalePrice')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def split_vals(a,n): return a[:n], a[n:]\n",
    "n_valid = 12000\n",
    "n_trn = len(df_trn)-n_valid\n",
    "X_train, X_valid = split_vals(df_trn, n_trn)\n",
    "y_train, y_valid = split_vals(y_trn, n_trn)\n",
    "raw_train, raw_valid = split_vals(df_raw, n_trn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x_sub = X_train[['YearMade', 'MachineHoursCurrentMeter']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.1 Basic Data Structure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class TreeEnsemble():\n",
    "    def __init__(self, x, y, n_trees, sample_sz, min_leaf=5):\n",
    "        np.random.seed(42)\n",
    "        self.x,self.y,self.sample_sz,self.min_leaf = x,y,sample_sz,min_leaf\n",
    "        self.trees = [self.create_tree() for i in range(n_trees)]\n",
    "    \n",
    "    def create_tree(self):\n",
    "        # get random indexes\n",
    "        rnd_idxs = np.random.permutation(len(self.y))[:self.sample_sz]\n",
    "        \n",
    "        # then return a decision tree based on a random dataset\n",
    "        return DecisionTree(self.x.iloc[rnd_idxs], self.y[rnd_idxs], min_leaf=self.min_leaf)\n",
    "        \n",
    "    def predict(self, x):\n",
    "        \n",
    "        # currently returning the mean of the prediction of all the \n",
    "        # trees\n",
    "        return np.mean([t.predict(x) for t in self.trees], axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.2 Decision Tree Version 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class DecisionTree():\n",
    "    def __init__(self, x, y, idxs=None, min_leaf=5):\n",
    "        self.x,self.y,self.idxs,self.min_leaf = x,y,idxs,min_leaf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Now we need to flesh out the decision Tree.\n",
    "\n",
    "We are going to write it as if everything exists\n",
    "\n",
    "### 2.1 Decision Tree Version 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "class DecisionTree():\n",
    "    def __init__(self, x, y, idxs=None, min_leaf=5):\n",
    "        \"\"\"\n",
    "        x : input features\n",
    "        y : target variable\n",
    "        idxs: keep track of what rows went where\n",
    "        \"\"\"\n",
    "        \n",
    "        # if its the first run, set to all\n",
    "        if idxs is None: idxs=np.arange(len(y))\n",
    "            \n",
    "        # store the data, and count rows and cols\n",
    "        self.x,self.y,self.idxs,self.min_leaf = x,y,idxs,min_leaf\n",
    "        self.n,self.c = len(idxs), x.shape[1]\n",
    "        \n",
    "        # every node in the tree will have the mean\n",
    "        # target variable\n",
    "        self.val = np.mean(y[idxs])\n",
    "        \n",
    "        # how effective was the split at this tree\n",
    "        # only if not a leaf node\n",
    "        self.score = float('inf')\n",
    "        \n",
    "        # having building root, which vari\n",
    "        # should we split and how much should we split on?\n",
    "        self.find_varsplit()\n",
    "        \n",
    "        \n",
    "    # This just does one decision; we'll make it recursive later\n",
    "    def find_varsplit(self):\n",
    "        \n",
    "        # go through each variable, and find a better split\n",
    "        # assuming all features are sent\n",
    "        for i in range(self.c): self.find_better_split(i)\n",
    "            \n",
    "    # We'll write this later!\n",
    "    def find_better_split(self, var_idx): pass\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Property means, dont need to call ()\n",
    "    will be calculated on the fly\n",
    "    \"\"\"\n",
    "    \n",
    "    @property\n",
    "    def split_name(self): return self.x.columns[self.var_idx]\n",
    "    \n",
    "    @property\n",
    "    def split_col(self): return self.x.values[self.idxs,self.var_idx]\n",
    "        \n",
    "    @property\n",
    "    def is_leaf(self): return self.score == float('inf')\n",
    "\n",
    "        \n",
    "    def __repr__(self):\n",
    "        \"\"\"\n",
    "        This is overloading the print function\n",
    "        so that everytime we print the decision tree\n",
    "        it will print some sort of data\n",
    "        \"\"\"\n",
    "        s = f'n: {self.n}; val:{self.val}'\n",
    "        if not self.is_leaf:\n",
    "            s += f'; score:{self.score}; split:{self.split}; var:{self.split_name}'\n",
    "        return s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "m = TreeEnsemble(X_train, y_train, n_trees=10, sample_sz=1000, min_leaf=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Test our __repr___ implementation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "m.trees[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Look at all the indexes in the tree (hint: all)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m.trees[0].idxs[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  2.2 Comparison to Sklearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['YearMade', 'MachineHoursCurrentMeter'], dtype='object')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ens = TreeEnsemble(x_sub, y_train, 1, 1000)\n",
    "tree = ens.trees[0]\n",
    "x_samp,y_samp = tree.x, tree.y\n",
    "x_samp.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Scikit Learn - 1 Tree, 1 Decision, No bootstrapping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"268pt\" height=\"163pt\"\n",
       " viewBox=\"0.00 0.00 267.86 163.40\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 159.4)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-159.4 263.864,-159.4 263.864,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.839216\" stroke=\"black\" points=\"127.23,-109.924 0.498952,-109.924 0.498952,-45.9237 127.23,-45.9237 127.23,-109.924\"/>\n",
       "<text text-anchor=\"start\" x=\"8.43164\" y=\"-94.7237\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1974.5</text>\n",
       "<text text-anchor=\"start\" x=\"32.8906\" y=\"-80.7237\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"19.8647\" y=\"-66.7237\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 1000</text>\n",
       "<text text-anchor=\"start\" x=\"25.5078\" y=\"-52.7237\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.08</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\n",
       "<polygon fill=\"none\" stroke=\"black\" points=\"260.363,-147.924 163.365,-147.924 163.365,-97.9237 260.363,-97.9237 260.363,-147.924\"/>\n",
       "<text text-anchor=\"start\" x=\"180.891\" y=\"-132.724\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.26</text>\n",
       "<text text-anchor=\"start\" x=\"171.365\" y=\"-118.724\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 159</text>\n",
       "<text text-anchor=\"start\" x=\"177.008\" y=\"-104.724\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.66</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M127.629,-97.2689C136.183,-99.9057 144.929,-102.601 153.384,-105.207\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"152.559,-108.616 163.147,-108.216 154.621,-101.926 152.559,-108.616\"/>\n",
       "<text text-anchor=\"middle\" x=\"141.046\" y=\"-115.703\" font-family=\"Times,serif\" font-size=\"14.00\">True</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"260.363,-57.9237 163.365,-57.9237 163.365,-7.92373 260.363,-7.92373 260.363,-57.9237\"/>\n",
       "<text text-anchor=\"start\" x=\"180.891\" y=\"-42.7237\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"171.365\" y=\"-28.7237\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 841</text>\n",
       "<text text-anchor=\"start\" x=\"173.508\" y=\"-14.7237\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.16</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\"><title>0&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M127.629,-58.5785C136.183,-55.9418 144.929,-53.2461 153.384,-50.6402\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"154.621,-53.9214 163.147,-47.6312 152.559,-47.232 154.621,-53.9214\"/>\n",
       "<text text-anchor=\"middle\" x=\"141.046\" y=\"-31.7446\" font-family=\"Times,serif\" font-size=\"14.00\">False</text>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x10fc40278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = RandomForestRegressor(n_estimators=1, max_depth=1, bootstrap=False)\n",
    "m.fit(x_samp, y_samp)\n",
    "draw_tree(m.estimators_[0], x_samp, precision=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### For our function - `find a better split`\n",
    "\n",
    "Given this subset of a data we are going to try to find a value to split on. We will find the split that minimizes the stddev of each group. Its the same as minimizing RMSE\n",
    "\n",
    "|YM|Price | group 1 | group 2|\n",
    "|--|--|------| -----|\n",
    "|3|1||\n",
    "|4|1||\n",
    "|1|0||\n",
    "|6|1||\n",
    "|4|0||\n",
    "|1|0||"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find Better Split v1\n",
    "\n",
    "Let's implement. Remember previously:\n",
    "```python\n",
    " # We'll write this later!\n",
    "    def find_better_split(self, var_idx): pass\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def find_better_split(self, var_idx):\n",
    "    \"\"\"\n",
    "    will get the indexes \n",
    "    \"\"\"\n",
    "    x,y = self.x.values[self.idxs,var_idx], self.y[self.idxs]\n",
    "    \n",
    "    # go through every single value\n",
    "    for i in range(1,self.n-1):\n",
    "        \n",
    "        # separate out the data, arrays of booleans\n",
    "        # since doing a comparative\n",
    "        lhs = x<=x[i]\n",
    "        rhs = x>x[i]\n",
    "        \n",
    "        # if there's no real split, then stop\n",
    "        if rhs.sum()==0: continue\n",
    "            \n",
    "        # otherwise calculate the two STDs\n",
    "        lhs_std = y[lhs].std()\n",
    "        rhs_std = y[rhs].std()\n",
    "        \n",
    "        #then average the score\n",
    "        curr_score = lhs_std*lhs.sum() + rhs_std*rhs.sum()\n",
    "        \n",
    "        # is this better than our score so far\n",
    "        # its intialized as inf.\n",
    "        # if its better, then we store\n",
    "        # variable, ID, SCORE, and what it is\n",
    "        if curr_score<self.score: \n",
    "            self.var_idx,self.score,self.split = var_idx,curr_score,x[i]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### From Above\n",
    "\n",
    "```python\n",
    "ens = TreeEnsemble(x_sub, y_train, 1, 1000)\n",
    "tree = ens.trees[0]\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "75.7 ms ± 1.01 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "n: 1000; val:10.079014121552744; score:681.0184057251435; split:3744.0; var:MachineHoursCurrentMeter"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%timeit find_better_split(tree,1)\n",
    "tree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test it out on a tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 1000; val:10.079014121552744; score:658.5510186055949; split:1974.0; var:YearMade"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "find_better_split(tree,0); tree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Baseline Computational Complexity? n^2\n",
    "\n",
    "- Consider loops\n",
    "- Sorts\n",
    "- comparisons\n",
    "\n",
    "```python\n",
    "    # will do n times\n",
    "    for i in range(1,self.n-1):\n",
    "        \n",
    "        # will also compare against all values so N again\n",
    "        lhs = x<=x[i]\n",
    "        rhs = x>x[i]\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Find Better Split : Speeding Things up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tree = TreeEnsemble(x_sub, y_train, 1, 1000).trees[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Variance\n",
    "\n",
    "$$ mean(xy) - mean(x)mean(y) $$\n",
    "\n",
    "Common language: sort the data, keep track of sum + count of the dataset and go through the dataset once. And if we have  sum + count, we can keep calculating the mean. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|YM|Price |\n",
    "|--|--|\n",
    "|1|0|\n",
    "|1|0|\n",
    "|3|1|\n",
    "|4|1|\n",
    "|4|0|\n",
    "|6|1|"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def std_agg(cnt, s1, s2): return math.sqrt((s2/cnt) - (s1/cnt)**2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1 Find Better Split v2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def std_agg(cnt, s1, s2): return math.sqrt((s2/cnt) - (s1/cnt)**2)\n",
    "\n",
    "def find_better_split(self, var_idx):\n",
    "    x,y = self.x.values[self.idxs,var_idx], self.y[self.idxs]\n",
    "    \n",
    "    sort_idx = np.argsort(x)\n",
    "    sort_y,sort_x = y[sort_idx], x[sort_idx]\n",
    "    rhs_cnt,rhs_sum,rhs_sum2 = self.n, sort_y.sum(), (sort_y**2).sum()\n",
    "    lhs_cnt,lhs_sum,lhs_sum2 = 0,0.,0.\n",
    "\n",
    "    for i in range(0,self.n-self.min_leaf-1):\n",
    "        xi,yi = sort_x[i],sort_y[i]\n",
    "        lhs_cnt += 1; rhs_cnt -= 1\n",
    "        lhs_sum += yi; rhs_sum -= yi\n",
    "        lhs_sum2 += yi**2; rhs_sum2 -= yi**2\n",
    "        if i<self.min_leaf or xi==sort_x[i+1]:\n",
    "            continue\n",
    "            \n",
    "        lhs_std = std_agg(lhs_cnt, lhs_sum, lhs_sum2)\n",
    "        rhs_std = std_agg(rhs_cnt, rhs_sum, rhs_sum2)\n",
    "        curr_score = lhs_std*lhs_cnt + rhs_std*rhs_cnt\n",
    "        if curr_score<self.score: \n",
    "            self.var_idx,self.score,self.split = var_idx,curr_score,xi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " "
     ]
    }
   ],
   "source": [
    "%prun find_better_split(tree,1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Example Output\n",
    "\n",
    "```bash\n",
    "         797 function calls (793 primitive calls) in 0.002 seconds\n",
    "\n",
    "   Ordered by: internal time\n",
    "\n",
    "   ncalls  tottime  percall  cumtime  percall filename:lineno(function)\n",
    "        1    0.001    0.001    0.002    0.002 <ipython-input-32-89db6b5dcd57>:3(find_better_split)\n",
    "      318    0.000    0.000    0.000    0.000 <ipython-input-32-89db6b5dcd57>:1(std_agg)\n",
    "        1    0.000    0.000    0.000    0.000 internals.py:3675(_interleave)\n",
    "        3    0.000    0.000    0.000    0.000 {method 'reduce' of 'numpy.ufunc' objects}\n",
    "        1    0.000    0.000    0.000    0.000 {method 'argsort' of 'numpy.ndarray' objects}\n",
    "      318    0.000    0.000    0.000    0.000 {built-in method math.sqrt}\n",
    "        1    0.000    0.000    0.002    0.002 {built-in method builtins.exec}\n",
    "        2    0.000    0.000    0.000    0.000 numerictypes.py:942(_can_coerce_all)\n",
    "        1    0.000    0.000    0.000    0.000 cast.py:1049(find_common_type)\n",
    "        1    0.000    0.000    0.000    0.000 generic.py:3661(_protect_consolidate)\n",
    "    32/28    0.000    0.000    0.000    0.000 {built-in method builtins.len}\n",
    " ...\n",
    " ```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2 Python Method for assigning functions, (assigning Find Better Split) \n",
    "\n",
    "```\n",
    "DecisionTree.find_better_split = find_better_split\n",
    "```\n",
    "\n",
    "**`DecisionTree.find_better_split`** - our method in our custom class\n",
    "\n",
    "**`find_better_split`** <- the v2 of the find_better_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "DecisionTree.find_better_split = find_better_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 1000; val:10.079014121552744; score:658.5510186055565; split:1974.0; var:YearMade"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree = TreeEnsemble(x_sub, y_train, 1, 1000).trees[0]; tree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Note this is the same as what we ran separately before (see a few blocks up)\n",
    "\n",
    "```python\n",
    "find_better_split(tree,0); tree\n",
    "\n",
    "n: 1000; val:10.079014121552744; score:658.5510186055949; split:1974.0; var:YearMade\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. How to do more than 1 level?\n",
    "\n",
    "### Back to Scikit Learn Comparison with 2 levels\n",
    "\n",
    "Now we will have to find a decision split in both child nodes to make the next level down."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"532pt\" height=\"322pt\"\n",
       " viewBox=\"0.00 0.00 532.44 322.40\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 318.4)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-318.4 528.441,-318.4 528.441,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.439216\" stroke=\"black\" points=\"126.807,-190.584 0.0761004,-190.584 0.0761004,-126.584 126.807,-126.584 126.807,-190.584\"/>\n",
       "<text text-anchor=\"start\" x=\"8.00879\" y=\"-175.384\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1974.5</text>\n",
       "<text text-anchor=\"start\" x=\"32.4678\" y=\"-161.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"19.4419\" y=\"-147.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 1000</text>\n",
       "<text text-anchor=\"start\" x=\"25.085\" y=\"-133.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.08</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.039216\" stroke=\"black\" points=\"391.519,-241.584 163.364,-241.584 163.364,-177.584 391.519,-177.584 391.519,-241.584\"/>\n",
       "<text text-anchor=\"start\" x=\"171.153\" y=\"-226.384\" font-family=\"Times,serif\" font-size=\"14.00\">MachineHoursCurrentMeter ≤ 2956.5</text>\n",
       "<text text-anchor=\"start\" x=\"246.468\" y=\"-212.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.26</text>\n",
       "<text text-anchor=\"start\" x=\"236.942\" y=\"-198.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 159</text>\n",
       "<text text-anchor=\"start\" x=\"242.585\" y=\"-184.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.66</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M126.902,-173.609C135.333,-175.638 144.209,-177.773 153.257,-179.95\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"152.588,-183.389 163.129,-182.325 154.225,-176.583 152.588,-183.389\"/>\n",
       "<text text-anchor=\"middle\" x=\"141.807\" y=\"-191.177\" font-family=\"Times,serif\" font-size=\"14.00\">True</text>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.513725\" stroke=\"black\" points=\"340.807,-140.584 214.076,-140.584 214.076,-76.5837 340.807,-76.5837 340.807,-140.584\"/>\n",
       "<text text-anchor=\"start\" x=\"222.009\" y=\"-125.384\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 2005.5</text>\n",
       "<text text-anchor=\"start\" x=\"246.468\" y=\"-111.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"236.942\" y=\"-97.3837\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 841</text>\n",
       "<text text-anchor=\"start\" x=\"239.085\" y=\"-83.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.16</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\"><title>0&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M126.902,-143.852C151.162,-138.131 179.106,-131.54 204.224,-125.616\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"205.132,-128.998 214.061,-123.296 203.525,-122.185 205.132,-128.998\"/>\n",
       "<text text-anchor=\"middle\" x=\"192.798\" y=\"-105.948\" font-family=\"Times,serif\" font-size=\"14.00\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\n",
       "<polygon fill=\"none\" stroke=\"black\" points=\"524.94,-308.584 427.942,-308.584 427.942,-258.584 524.94,-258.584 524.94,-308.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-293.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.23</text>\n",
       "<text text-anchor=\"start\" x=\"435.942\" y=\"-279.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 150</text>\n",
       "<text text-anchor=\"start\" x=\"441.585\" y=\"-265.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.62</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M363.703,-241.611C382.044,-248.5 401.01,-255.625 417.992,-262.004\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"417.15,-265.426 427.742,-265.666 419.612,-258.873 417.15,-265.426\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.698039\" stroke=\"black\" points=\"522.655,-225.584 430.228,-225.584 430.228,-175.584 522.655,-175.584 522.655,-225.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-210.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.23</text>\n",
       "<text text-anchor=\"start\" x=\"442.942\" y=\"-196.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 9</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-182.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.35</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\"><title>1&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M391.61,-204.414C401.348,-203.969 410.89,-203.533 419.845,-203.124\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"420.233,-206.61 430.063,-202.657 419.914,-199.617 420.233,-206.61\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.494118\" stroke=\"black\" points=\"524.94,-138.584 427.942,-138.584 427.942,-88.5837 524.94,-88.5837 524.94,-138.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-123.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.46</text>\n",
       "<text text-anchor=\"start\" x=\"435.942\" y=\"-109.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 813</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-95.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.14</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\"><title>4&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M341.044,-110.173C365.591,-110.796 393.584,-111.506 417.644,-112.117\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"417.608,-115.617 427.694,-112.372 417.786,-108.619 417.608,-115.617\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"522.655,-55.5837 430.228,-55.5837 430.228,-5.58366 522.655,-5.58366 522.655,-55.5837\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-40.3837\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.38</text>\n",
       "<text text-anchor=\"start\" x=\"439.442\" y=\"-26.3837\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 28</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-12.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.67</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\"><title>4&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M341.044,-83.7968C366.606,-73.6759 395.904,-62.0756 420.608,-52.2944\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"422.174,-55.4387 430.183,-48.5031 419.597,-48.9303 422.174,-55.4387\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x1c1bea1ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = RandomForestRegressor(n_estimators=1, max_depth=2, bootstrap=False)\n",
    "m.fit(x_samp, y_samp)\n",
    "draw_tree(m.estimators_[0], x_samp, precision=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.1 Find_varsplit version 2 (improvements)\n",
    " \n",
    "Old version:\n",
    "\n",
    "```python\n",
    "  # This just does one decision; we'll make it recursive later\n",
    "    def find_varsplit(self):\n",
    "        for i in range(self.c): self.find_better_split(i)\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def find_varsplit(self):\n",
    "    \n",
    "    # is there something better\n",
    "    for i in range(self.c): self.find_better_split(i)\n",
    "    if self.is_leaf: return\n",
    "    x = self.split_col\n",
    "    \n",
    "    # gets the boolean array, but gets the indexes\n",
    "    lhs = np.nonzero(x<=self.split)[0]\n",
    "    rhs = np.nonzero(x>self.split)[0]\n",
    "    \n",
    "    # now that we have the indexes we can create a decision tree\n",
    "    self.lhs = DecisionTree(self.x, self.y, self.idxs[lhs])\n",
    "    self.rhs = DecisionTree(self.x, self.y, self.idxs[rhs])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Assign it to our class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "DecisionTree.find_varsplit = find_varsplit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 1000; val:10.079014121552744; score:658.5510186055565; split:1974.0; var:YearMade"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree = TreeEnsemble(x_sub, y_train, 1, 1000).trees[0]; tree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.2 Checking Answers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Lets look at the left hand and right hand side"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 159; val:9.660892662981706; score:76.82696888346362; split:2800.0; var:MachineHoursCurrentMeter"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.lhs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 841; val:10.158064432982941; score:571.4803525045031; split:2005.0; var:YearMade"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.rhs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### What about level2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 150; val:9.619280538108496; score:71.15906938383463; split:1000.0; var:YearMade"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.lhs.lhs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 9; val:10.354428077535193"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.lhs.rhs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 813; val:10.140438373816052; score:550.1806493212129; split:1994.0; var:YearMade"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.rhs.lhs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "n: 28; val:10.669849650935848; score:16.460871194578985; split:895.0; var:MachineHoursCurrentMeter"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tree.rhs.rhs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Again compare to Sklearn (look at samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"532pt\" height=\"322pt\"\n",
       " viewBox=\"0.00 0.00 532.44 322.40\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 318.4)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-318.4 528.441,-318.4 528.441,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.439216\" stroke=\"black\" points=\"126.807,-190.584 0.0761004,-190.584 0.0761004,-126.584 126.807,-126.584 126.807,-190.584\"/>\n",
       "<text text-anchor=\"start\" x=\"8.00879\" y=\"-175.384\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1974.5</text>\n",
       "<text text-anchor=\"start\" x=\"32.4678\" y=\"-161.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"19.4419\" y=\"-147.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 1000</text>\n",
       "<text text-anchor=\"start\" x=\"25.085\" y=\"-133.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.08</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.039216\" stroke=\"black\" points=\"391.519,-241.584 163.364,-241.584 163.364,-177.584 391.519,-177.584 391.519,-241.584\"/>\n",
       "<text text-anchor=\"start\" x=\"171.153\" y=\"-226.384\" font-family=\"Times,serif\" font-size=\"14.00\">MachineHoursCurrentMeter ≤ 2956.5</text>\n",
       "<text text-anchor=\"start\" x=\"246.468\" y=\"-212.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.26</text>\n",
       "<text text-anchor=\"start\" x=\"236.942\" y=\"-198.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 159</text>\n",
       "<text text-anchor=\"start\" x=\"242.585\" y=\"-184.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.66</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M126.902,-173.609C135.333,-175.638 144.209,-177.773 153.257,-179.95\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"152.588,-183.389 163.129,-182.325 154.225,-176.583 152.588,-183.389\"/>\n",
       "<text text-anchor=\"middle\" x=\"141.807\" y=\"-191.177\" font-family=\"Times,serif\" font-size=\"14.00\">True</text>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.513725\" stroke=\"black\" points=\"340.807,-140.584 214.076,-140.584 214.076,-76.5837 340.807,-76.5837 340.807,-140.584\"/>\n",
       "<text text-anchor=\"start\" x=\"222.009\" y=\"-125.384\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 2005.5</text>\n",
       "<text text-anchor=\"start\" x=\"246.468\" y=\"-111.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"236.942\" y=\"-97.3837\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 841</text>\n",
       "<text text-anchor=\"start\" x=\"239.085\" y=\"-83.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.16</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\"><title>0&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M126.902,-143.852C151.162,-138.131 179.106,-131.54 204.224,-125.616\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"205.132,-128.998 214.061,-123.296 203.525,-122.185 205.132,-128.998\"/>\n",
       "<text text-anchor=\"middle\" x=\"192.798\" y=\"-105.948\" font-family=\"Times,serif\" font-size=\"14.00\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\n",
       "<polygon fill=\"none\" stroke=\"black\" points=\"524.94,-308.584 427.942,-308.584 427.942,-258.584 524.94,-258.584 524.94,-308.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-293.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.23</text>\n",
       "<text text-anchor=\"start\" x=\"435.942\" y=\"-279.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 150</text>\n",
       "<text text-anchor=\"start\" x=\"441.585\" y=\"-265.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.62</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M363.703,-241.611C382.044,-248.5 401.01,-255.625 417.992,-262.004\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"417.15,-265.426 427.742,-265.666 419.612,-258.873 417.15,-265.426\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.698039\" stroke=\"black\" points=\"522.655,-225.584 430.228,-225.584 430.228,-175.584 522.655,-175.584 522.655,-225.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-210.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.23</text>\n",
       "<text text-anchor=\"start\" x=\"442.942\" y=\"-196.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 9</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-182.384\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.35</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\"><title>1&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M391.61,-204.414C401.348,-203.969 410.89,-203.533 419.845,-203.124\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"420.233,-206.61 430.063,-202.657 419.914,-199.617 420.233,-206.61\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.494118\" stroke=\"black\" points=\"524.94,-138.584 427.942,-138.584 427.942,-88.5837 524.94,-88.5837 524.94,-138.584\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-123.384\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.46</text>\n",
       "<text text-anchor=\"start\" x=\"435.942\" y=\"-109.384\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 813</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-95.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.14</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\"><title>4&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M341.044,-110.173C365.591,-110.796 393.584,-111.506 417.644,-112.117\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"417.608,-115.617 427.694,-112.372 417.786,-108.619 417.608,-115.617\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"522.655,-55.5837 430.228,-55.5837 430.228,-5.58366 522.655,-5.58366 522.655,-55.5837\"/>\n",
       "<text text-anchor=\"start\" x=\"445.468\" y=\"-40.3837\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.38</text>\n",
       "<text text-anchor=\"start\" x=\"439.442\" y=\"-26.3837\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 28</text>\n",
       "<text text-anchor=\"start\" x=\"438.085\" y=\"-12.3837\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.67</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\"><title>4&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M341.044,-83.7968C366.606,-73.6759 395.904,-62.0756 420.608,-52.2944\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"422.174,-55.4387 430.183,-48.5031 419.597,-48.9303 422.174,-55.4387\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x1c1bea1cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_tree(m.estimators_[0], x_samp, precision=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Writing our RF Predictor\n",
    "\n",
    "Lets look through a basic prototype of the prediction function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cols = ['MachineID', 'YearMade', 'MachineHoursCurrentMeter', 'ProductSize', 'Enclosure',\n",
    "        'Coupler_System', 'saleYear']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 287 ms, sys: 6.72 ms, total: 294 ms\n",
      "Wall time: 293 ms\n"
     ]
    }
   ],
   "source": [
    "%time tree = TreeEnsemble(X_train[cols], y_train, 1, 1000).trees[0]\n",
    "x_samp,y_samp = tree.x, tree.y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The Sk learn version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"504pt\" height=\"454pt\"\n",
       " viewBox=\"0.00 0.00 504.00 454.08\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(0.828615 0.828615) rotate(0) translate(4 544)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-544 604.244,-544 604.244,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.729412\" stroke=\"black\" points=\"141.76,-301.665 0.728663,-301.665 0.728663,-237.665 141.76,-237.665 141.76,-301.665\"/>\n",
       "<text text-anchor=\"start\" x=\"8.48682\" y=\"-286.465\" font-family=\"Times,serif\" font-size=\"14.00\">Coupler_System ≤ 0.5</text>\n",
       "<text text-anchor=\"start\" x=\"40.2705\" y=\"-272.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.47</text>\n",
       "<text text-anchor=\"start\" x=\"27.2446\" y=\"-258.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 1000</text>\n",
       "<text text-anchor=\"start\" x=\"32.8877\" y=\"-244.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.08</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.796078\" stroke=\"black\" points=\"304.609,-354.665 177.879,-354.665 177.879,-290.665 304.609,-290.665 304.609,-354.665\"/>\n",
       "<text text-anchor=\"start\" x=\"185.812\" y=\"-339.465\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1980.5</text>\n",
       "<text text-anchor=\"start\" x=\"210.271\" y=\"-325.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.41</text>\n",
       "<text text-anchor=\"start\" x=\"200.745\" y=\"-311.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 898</text>\n",
       "<text text-anchor=\"start\" x=\"202.888\" y=\"-297.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.18</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M142.032,-291.682C150.612,-294.389 159.397,-297.16 168.013,-299.878\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"167.059,-303.247 177.649,-302.918 169.165,-296.572 167.059,-303.247\"/>\n",
       "<text text-anchor=\"middle\" x=\"155.472\" y=\"-310.258\" font-family=\"Times,serif\" font-size=\"14.00\">True</text>\n",
       "</g>\n",
       "<!-- 8 -->\n",
       "<g id=\"node9\" class=\"node\"><title>8</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.129412\" stroke=\"black\" points=\"304.609,-249.665 177.879,-249.665 177.879,-185.665 304.609,-185.665 304.609,-249.665\"/>\n",
       "<text text-anchor=\"start\" x=\"185.812\" y=\"-234.465\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1998.5</text>\n",
       "<text text-anchor=\"start\" x=\"210.271\" y=\"-220.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.12</text>\n",
       "<text text-anchor=\"start\" x=\"200.745\" y=\"-206.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 102</text>\n",
       "<text text-anchor=\"start\" x=\"206.388\" y=\"-192.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.17</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;8 -->\n",
       "<g id=\"edge8\" class=\"edge\"><title>0&#45;&gt;8</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M142.032,-248.064C150.612,-245.409 159.397,-242.69 168.013,-240.023\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"169.131,-243.341 177.649,-237.04 167.061,-236.654 169.131,-243.341\"/>\n",
       "<text text-anchor=\"middle\" x=\"155.535\" y=\"-221.18\" font-family=\"Times,serif\" font-size=\"14.00\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.568627\" stroke=\"black\" points=\"467.609,-476.665 340.879,-476.665 340.879,-412.665 467.609,-412.665 467.609,-476.665\"/>\n",
       "<text text-anchor=\"start\" x=\"348.812\" y=\"-461.465\" font-family=\"Times,serif\" font-size=\"14.00\">YearMade ≤ 1974.5</text>\n",
       "<text text-anchor=\"start\" x=\"373.271\" y=\"-447.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.28</text>\n",
       "<text text-anchor=\"start\" x=\"363.745\" y=\"-433.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 226</text>\n",
       "<text text-anchor=\"start\" x=\"369.388\" y=\"-419.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.83</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M284.545,-354.719C305.446,-370.558 330.82,-389.785 352.734,-406.39\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"350.863,-409.364 360.947,-412.614 355.09,-403.785 350.863,-409.364\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.874510\" stroke=\"black\" points=\"462.083,-354.665 346.405,-354.665 346.405,-290.665 462.083,-290.665 462.083,-354.665\"/>\n",
       "<text text-anchor=\"start\" x=\"354.325\" y=\"-339.465\" font-family=\"Times,serif\" font-size=\"14.00\">ProductSize ≤ 1.5</text>\n",
       "<text text-anchor=\"start\" x=\"373.271\" y=\"-325.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.39</text>\n",
       "<text text-anchor=\"start\" x=\"363.745\" y=\"-311.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 672</text>\n",
       "<text text-anchor=\"start\" x=\"369.388\" y=\"-297.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.3</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\"><title>1&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M304.952,-322.665C315.15,-322.665 325.749,-322.665 336.049,-322.665\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"336.223,-326.165 346.223,-322.665 336.223,-319.165 336.223,-326.165\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.490196\" stroke=\"black\" points=\"600.743,-539.665 503.745,-539.665 503.745,-489.665 600.743,-489.665 600.743,-539.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-524.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.25</text>\n",
       "<text text-anchor=\"start\" x=\"511.745\" y=\"-510.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 143</text>\n",
       "<text text-anchor=\"start\" x=\"517.388\" y=\"-496.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.72</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\"><title>2&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M468.009,-474.758C476.749,-478.949 485.69,-483.235 494.315,-487.371\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"492.996,-490.62 503.527,-491.787 496.023,-484.308 492.996,-490.62\"/>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.694118\" stroke=\"black\" points=\"598.457,-469.665 506.031,-469.665 506.031,-419.665 598.457,-419.665 598.457,-469.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-454.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.29</text>\n",
       "<text text-anchor=\"start\" x=\"515.245\" y=\"-440.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 83</text>\n",
       "<text text-anchor=\"start\" x=\"513.888\" y=\"-426.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.03</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\"><title>2&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M468.009,-444.665C477.218,-444.665 486.649,-444.665 495.699,-444.665\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"495.724,-448.165 505.724,-444.665 495.724,-441.165 495.724,-448.165\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.752941\" stroke=\"black\" points=\"600.743,-399.665 503.745,-399.665 503.745,-349.665 600.743,-349.665 600.743,-399.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-384.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.29</text>\n",
       "<text text-anchor=\"start\" x=\"511.745\" y=\"-370.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 341</text>\n",
       "<text text-anchor=\"start\" x=\"514.144\" y=\"-356.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.11</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\"><title>5&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.11,-342.919C472.531,-346.631 483.414,-350.507 493.85,-354.224\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"492.689,-357.525 503.283,-357.584 495.037,-350.931 492.689,-357.525\"/>\n",
       "</g>\n",
       "<!-- 7 -->\n",
       "<g id=\"node8\" class=\"node\"><title>7</title>\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"600.743,-329.665 503.745,-329.665 503.745,-279.665 600.743,-279.665 600.743,-329.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-314.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.43</text>\n",
       "<text text-anchor=\"start\" x=\"511.745\" y=\"-300.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 331</text>\n",
       "<text text-anchor=\"start\" x=\"513.888\" y=\"-286.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 10.49</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;7 -->\n",
       "<g id=\"edge7\" class=\"edge\"><title>5&#45;&gt;7</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.11,-315.655C472.318,-314.396 482.97,-313.083 493.211,-311.82\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"493.787,-315.276 503.283,-310.578 492.93,-308.328 493.787,-315.276\"/>\n",
       "</g>\n",
       "<!-- 9 -->\n",
       "<g id=\"node10\" class=\"node\"><title>9</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.031373\" stroke=\"black\" points=\"462.499,-249.665 345.989,-249.665 345.989,-185.665 462.499,-185.665 462.499,-249.665\"/>\n",
       "<text text-anchor=\"start\" x=\"353.867\" y=\"-234.465\" font-family=\"Times,serif\" font-size=\"14.00\">saleYear ≤ 1995.5</text>\n",
       "<text text-anchor=\"start\" x=\"376.771\" y=\"-220.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.1</text>\n",
       "<text text-anchor=\"start\" x=\"367.245\" y=\"-206.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 49</text>\n",
       "<text text-anchor=\"start\" x=\"369.388\" y=\"-192.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.02</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;9 -->\n",
       "<g id=\"edge9\" class=\"edge\"><title>8&#45;&gt;9</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M304.952,-217.665C315.084,-217.665 325.611,-217.665 335.848,-217.665\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"335.963,-221.165 345.963,-217.665 335.963,-214.165 335.963,-221.165\"/>\n",
       "</g>\n",
       "<!-- 12 -->\n",
       "<g id=\"node13\" class=\"node\"><title>12</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.219608\" stroke=\"black\" points=\"462.499,-127.665 345.989,-127.665 345.989,-63.6654 462.499,-63.6654 462.499,-127.665\"/>\n",
       "<text text-anchor=\"start\" x=\"353.867\" y=\"-112.465\" font-family=\"Times,serif\" font-size=\"14.00\">saleYear ≤ 2007.5</text>\n",
       "<text text-anchor=\"start\" x=\"376.771\" y=\"-98.4654\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.1</text>\n",
       "<text text-anchor=\"start\" x=\"367.245\" y=\"-84.4654\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 53</text>\n",
       "<text text-anchor=\"start\" x=\"369.388\" y=\"-70.4654\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.31</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;12 -->\n",
       "<g id=\"edge12\" class=\"edge\"><title>8&#45;&gt;12</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M284.545,-185.611C305.446,-169.773 330.82,-150.546 352.734,-133.94\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"355.09,-136.546 360.947,-127.717 350.863,-130.967 355.09,-136.546\"/>\n",
       "</g>\n",
       "<!-- 10 -->\n",
       "<g id=\"node11\" class=\"node\"><title>10</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.317647\" stroke=\"black\" points=\"594.958,-259.665 509.531,-259.665 509.531,-209.665 594.958,-209.665 594.958,-259.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-244.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.03</text>\n",
       "<text text-anchor=\"start\" x=\"518.745\" y=\"-230.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 5</text>\n",
       "<text text-anchor=\"start\" x=\"517.388\" y=\"-216.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.46</text>\n",
       "</g>\n",
       "<!-- 9&#45;&gt;10 -->\n",
       "<g id=\"edge10\" class=\"edge\"><title>9&#45;&gt;10</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.528,-224.335C474.639,-225.746 487.355,-227.226 499.257,-228.612\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"499.133,-232.121 509.471,-229.801 499.942,-225.168 499.133,-232.121\"/>\n",
       "</g>\n",
       "<!-- 11 -->\n",
       "<g id=\"node12\" class=\"node\"><title>11</title>\n",
       "<polygon fill=\"none\" stroke=\"black\" points=\"597.243,-190.665 507.245,-190.665 507.245,-140.665 597.243,-140.665 597.243,-190.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-175.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.08</text>\n",
       "<text text-anchor=\"start\" x=\"515.245\" y=\"-161.465\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 44</text>\n",
       "<text text-anchor=\"start\" x=\"517.388\" y=\"-147.465\" font-family=\"Times,serif\" font-size=\"14.00\">value = 8.97</text>\n",
       "</g>\n",
       "<!-- 9&#45;&gt;11 -->\n",
       "<g id=\"edge11\" class=\"edge\"><title>9&#45;&gt;11</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.528,-197.263C474.137,-193.128 486.302,-188.795 497.773,-184.71\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"498.972,-187.998 507.218,-181.346 496.623,-181.404 498.972,-187.998\"/>\n",
       "</g>\n",
       "<!-- 13 -->\n",
       "<g id=\"node14\" class=\"node\"><title>13</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.333333\" stroke=\"black\" points=\"597.243,-120.665 507.245,-120.665 507.245,-70.6654 597.243,-70.6654 597.243,-120.665\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-105.465\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.08</text>\n",
       "<text text-anchor=\"start\" x=\"515.245\" y=\"-91.4654\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 22</text>\n",
       "<text text-anchor=\"start\" x=\"517.388\" y=\"-77.4654\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.48</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;13 -->\n",
       "<g id=\"edge13\" class=\"edge\"><title>12&#45;&gt;13</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.528,-95.6654C473.905,-95.6654 485.816,-95.6654 497.084,-95.6654\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"497.218,-99.1655 507.218,-95.6654 497.218,-92.1655 497.218,-99.1655\"/>\n",
       "</g>\n",
       "<!-- 14 -->\n",
       "<g id=\"node15\" class=\"node\"><title>14</title>\n",
       "<polygon fill=\"#e58139\" fill-opacity=\"0.141176\" stroke=\"black\" points=\"597.243,-50.6654 507.245,-50.6654 507.245,-0.665399 597.243,-0.665399 597.243,-50.6654\"/>\n",
       "<text text-anchor=\"start\" x=\"521.271\" y=\"-35.4654\" font-family=\"Times,serif\" font-size=\"14.00\">mse = 0.08</text>\n",
       "<text text-anchor=\"start\" x=\"515.245\" y=\"-21.4654\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 31</text>\n",
       "<text text-anchor=\"start\" x=\"517.388\" y=\"-7.4654\" font-family=\"Times,serif\" font-size=\"14.00\">value = 9.19</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;14 -->\n",
       "<g id=\"edge14\" class=\"edge\"><title>12&#45;&gt;14</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M462.528,-68.2005C474.253,-62.579 486.546,-56.6852 498.117,-51.1372\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"499.714,-54.2531 507.218,-46.7738 496.688,-47.9411 499.714,-54.2531\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x1c1cebf908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = RandomForestRegressor(n_estimators=1, max_depth=3, bootstrap=False)\n",
    "m.fit(x_samp, y_samp)\n",
    "draw_tree(m.estimators_[0], x_samp, precision=2, ratio=0.9, size=7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def predict(self, x): return np.array([self.predict_row(xi) for xi in x])\n",
    "    \n",
    "DecisionTree.predict = predict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Inline If functions\n",
    "\n",
    "```python\n",
    "if x > 1:\n",
    "    x = \"greater\"\n",
    "else:\n",
    "    x =\"lesser\"\n",
    "```\n",
    "\n",
    "Can be written as the following:\n",
    "\n",
    "```python\n",
    "x = \"greater\" if x > 1 else \"lesser\"\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def predict_row(self, xi):\n",
    "    \n",
    "    #if leaf return\n",
    "    if self.is_leaf: return self.val\n",
    "    \n",
    "    # otherwise return lhs or right hand side\n",
    "    # this is an exampel of inline if value\n",
    "    t = self.lhs if xi[self.var_idx]<=self.split else self.rhs\n",
    "    return t.predict_row(xi)\n",
    "\n",
    "DecisionTree.predict_row = predict_row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 89 ms, sys: 2.18 ms, total: 91.2 ms\n",
      "Wall time: 90.6 ms\n"
     ]
    }
   ],
   "source": [
    "%time preds = tree.predict(X_valid[cols].values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Looking at scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1c1e4ca438>"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAD8CAYAAABw1c+bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUBwwlOJXKBSKA8b/B4gpiakLpeoyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c1cec4630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(preds, y_valid, alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.50371522136882341"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.r2_score(preds, y_valid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare to the Sklearn Plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3.38 ms, sys: 632 µs, total: 4.02 ms\n",
      "Wall time: 3.44 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1c1e669240>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAD8CAYAAABw1c+bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzXgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117f5d4e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = RandomForestRegressor(n_estimators=1, min_samples_leaf=5, bootstrap=False)\n",
    "%time m.fit(x_samp, y_samp)\n",
    "preds = m.predict(X_valid[cols].values)\n",
    "plt.scatter(preds, y_valid, alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.47541053100694763"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.r2_score(preds, y_valid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Assembling the Final Version\n",
    "\n",
    "And below we cut and paste all the necessary functions and collect it in two neat classes as listed down below"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class TreeEnsemble():\n",
    "    def __init__(self, x, y, n_trees, sample_sz, min_leaf=5):\n",
    "        np.random.seed(42)\n",
    "        self.x,self.y,self.sample_sz,self.min_leaf = x,y,sample_sz,min_leaf\n",
    "        self.trees = [self.create_tree() for i in range(n_trees)]\n",
    "\n",
    "    def create_tree(self):\n",
    "        idxs = np.random.permutation(len(self.y))[:self.sample_sz]\n",
    "        return DecisionTree(self.x.iloc[idxs], self.y[idxs], \n",
    "                    idxs=np.array(range(self.sample_sz)), min_leaf=self.min_leaf)\n",
    "        \n",
    "    def predict(self, x):\n",
    "        return np.mean([t.predict(x) for t in self.trees], axis=0)\n",
    "\n",
    "def std_agg(cnt, s1, s2): return math.sqrt((s2/cnt) - (s1/cnt)**2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class DecisionTree():\n",
    "    def __init__(self, x, y, idxs, min_leaf=5):\n",
    "        self.x,self.y,self.idxs,self.min_leaf = x,y,idxs,min_leaf\n",
    "        self.n,self.c = len(idxs), x.shape[1]\n",
    "        self.val = np.mean(y[idxs])\n",
    "        self.score = float('inf')\n",
    "        self.find_varsplit()\n",
    "        \n",
    "    def find_varsplit(self):\n",
    "        for i in range(self.c): self.find_better_split(i)\n",
    "        if self.score == float('inf'): return\n",
    "        x = self.split_col\n",
    "        lhs = np.nonzero(x<=self.split)[0]\n",
    "        rhs = np.nonzero(x>self.split)[0]\n",
    "        self.lhs = DecisionTree(self.x, self.y, self.idxs[lhs])\n",
    "        self.rhs = DecisionTree(self.x, self.y, self.idxs[rhs])\n",
    "\n",
    "    def find_better_split(self, var_idx):\n",
    "        x,y = self.x.values[self.idxs,var_idx], self.y[self.idxs]\n",
    "        sort_idx = np.argsort(x)\n",
    "        sort_y,sort_x = y[sort_idx], x[sort_idx]\n",
    "        rhs_cnt,rhs_sum,rhs_sum2 = self.n, sort_y.sum(), (sort_y**2).sum()\n",
    "        lhs_cnt,lhs_sum,lhs_sum2 = 0,0.,0.\n",
    "\n",
    "        for i in range(0,self.n-self.min_leaf-1):\n",
    "            xi,yi = sort_x[i],sort_y[i]\n",
    "            lhs_cnt += 1; rhs_cnt -= 1\n",
    "            lhs_sum += yi; rhs_sum -= yi\n",
    "            lhs_sum2 += yi**2; rhs_sum2 -= yi**2\n",
    "            if i<self.min_leaf or xi==sort_x[i+1]:\n",
    "                continue\n",
    "\n",
    "            lhs_std = std_agg(lhs_cnt, lhs_sum, lhs_sum2)\n",
    "            rhs_std = std_agg(rhs_cnt, rhs_sum, rhs_sum2)\n",
    "            curr_score = lhs_std*lhs_cnt + rhs_std*rhs_cnt\n",
    "            if curr_score<self.score: \n",
    "                self.var_idx,self.score,self.split = var_idx,curr_score,xi\n",
    "\n",
    "    @property\n",
    "    def split_name(self): return self.x.columns[self.var_idx]\n",
    "    \n",
    "    @property\n",
    "    def split_col(self): return self.x.values[self.idxs,self.var_idx]\n",
    "\n",
    "    @property\n",
    "    def is_leaf(self): return self.score == float('inf')\n",
    "    \n",
    "    def __repr__(self):\n",
    "        s = f'n: {self.n}; val:{self.val}'\n",
    "        if not self.is_leaf:\n",
    "            s += f'; score:{self.score}; split:{self.split}; var:{self.split_name}'\n",
    "        return s\n",
    "\n",
    "    def predict(self, x):\n",
    "        return np.array([self.predict_row(xi) for xi in x])\n",
    "\n",
    "    def predict_row(self, xi):\n",
    "        if self.is_leaf: return self.val\n",
    "        t = self.lhs if xi[self.var_idx]<=self.split else self.rhs\n",
    "        return t.predict_row(xi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: Cython in /Users/tlee010/anaconda/envs/dsi/lib/python2.7/site-packages\r\n"
     ]
    }
   ],
   "source": [
    "!pip install Cython"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fetching package metadata ...........\n",
      "Solving package specifications: .\n",
      "\n",
      "Package plan for installation in environment /Users/tlee010/anaconda/envs/dsi:\n",
      "\n",
      "The following packages will be UPDATED:\n",
      "\n",
      "    cython: 0.24-py27_0 --> 0.26.1-py27h6a053f9_0\n",
      "\n",
      "Proceed ([y]/n)? ^C\n"
     ]
    }
   ],
   "source": [
    "!conda install Cython"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Cython - more speed improvements"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A super set of python"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%load_ext Cython"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Pure Python"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def fib1(n):\n",
    "    a, b = 0,1\n",
    "    while b < n:\n",
    "        a,b = b, a +b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "308 ns ± 8.24 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fib1(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Cython Wrapper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%%cython\n",
    "def fib2(n):\n",
    "    a, b = 0,1\n",
    "    while b < n:\n",
    "        a,b = b, a +b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "306 ns ± 6.26 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fib2(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Cython version with Typing Specified"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%cython\n",
    "def fib3(int n):\n",
    "    cdef int b = 1\n",
    "    cdef int a = 0\n",
    "    cdef int t = 0\n",
    "    while b < n:\n",
    "        t = a\n",
    "        a = b\n",
    "        b = t + b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "51.4 ns ± 0.903 ns per loop (mean ± std. dev. of 7 runs, 10000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fib3(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Homework\n",
    "\n",
    "#### Confidence Based Tree Variance\n",
    "#### Feature Importance\n",
    "#### Partial Dependence\n",
    "#### Tree Interpreter\n",
    "#### Bonus : Interaction Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [fastai]",
   "language": "python",
   "name": "Python [fastai]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}