state-classfication_safety-forecasting.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"'''\nimport pandas as pd\nimport numpy as np\nimport glob\n\npath = r'D://Research/DIT4BEARs/Road Measurements/Internship data/All' # use your path\nall_files = glob.glob(path + \"/*.csv\")\n\nli = []\n\nfor filename in all_files:\n    df = pd.read_csv(filename, index_col=None, header=0)\n    li.append(df)\n\nframe = pd.concat(li, axis=0, ignore_index=True)\n\nall_files\n'''","metadata":{"id":"-peMWWI6NU6P","outputId":"3c7191e8-43c4-483c-c7e3-d65f81642f79","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"'''\nframe.head()\n'''","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"'''\nframe.shape\n'''","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"'''\nframe.to_csv('D://Research/DIT4BEARs/Road Measurements/Internship data/UiTdata.csv')\n'''","metadata":{"id":"O920hIjJV_uQ"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Import Important libraries\n\nimport pandas as pd\nimport numpy as np\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n%matplotlib inline\nfrom sklearn.linear_model import LinearRegression\nfrom sklearn.model_selection import train_test_split\nfrom sklearn import linear_model\nfrom sklearn.linear_model import LinearRegression\nimport plotly.express as px\nfrom fbprophet.plot import plot_plotly, plot_components_plotly\nfrom fbprophet import Prophet\nfrom statsmodels.tsa.seasonal import seasonal_decompose as sd\nimport plotly.graph_objects as go\nimport statsmodels.api as sm\nimport statsmodels.tsa.api as smt\nimport missingno as msno\nfrom sklearn.model_selection import TimeSeriesSplit, train_test_split\nfrom sklearn.model_selection import cross_val_score\nfrom sklearn.preprocessing import StandardScaler\n\nfrom sklearn.ensemble import RandomForestRegressor\nfrom datetime import timedelta\nimport warnings\nwarnings.filterwarnings(\"ignore\")","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:19:42.20674Z","iopub.execute_input":"2021-05-20T18:19:42.207099Z","iopub.status.idle":"2021-05-20T18:19:46.181842Z","shell.execute_reply.started":"2021-05-20T18:19:42.20707Z","shell.execute_reply":"2021-05-20T18:19:46.180763Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"f = pd.read_csv('../input/road-dataset/UiTdata.csv')\nf.head()","metadata":{"id":"v5Efnj8QfbXF","outputId":"c50c10de-1212-4c25-f016-c7e184b65f8c","execution":{"iopub.status.busy":"2021-05-20T19:00:59.047851Z","iopub.execute_input":"2021-05-20T19:00:59.048439Z","iopub.status.idle":"2021-05-20T19:01:08.340415Z","shell.execute_reply.started":"2021-05-20T19:00:59.048403Z","shell.execute_reply":"2021-05-20T19:01:08.339244Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Data fields\n\nDate - The date for recording (YYYY-MM-DD)\\\nTime - The time when recording at a given date (HH:MM:SS.6f)\\\nS1 - Measurement signal 1 (not to be used)\\\nS2 - Measurement signal 2 (not to be used)\\\nS3 - Measurement signal 3 (not to be used)\\\nFriction - Measure friction value between 0.1-0.81\\\nState - Values of 1-6 for: dry, moist, wet, icy, snowy, slushy\\\nTa - Temperature in the air\\\nS7 - N/A\\\nTsurf - Temperature on the measured surface\\\nS9 - N/A\\\nS10 - N/A\\\nS11 -N/A\\\nWater - Water layer thicknes in 0-3 mm\\\nHeight - Above sealevel in m\\\nTdew - The dewpoint temperature (not used in every aspect)\\\nFriction2 - Another calculation of friction (not used)\\\nDistance - Driven distance in m since startup of a session\\\nSerial(sensor RCM411) - Serial number of the current sensor ","metadata":{"id":"OjwrGQ_A3gr5"}},{"cell_type":"code","source":"f.shape","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:19:58.556064Z","iopub.execute_input":"2021-05-20T18:19:58.556546Z","iopub.status.idle":"2021-05-20T18:19:58.563265Z","shell.execute_reply.started":"2021-05-20T18:19:58.55648Z","shell.execute_reply":"2021-05-20T18:19:58.562168Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"f.drop(columns=['Unnamed: 0','S1','S2','S3','S7','S9','S10','S11','Tdew','Friction2','Serial(RCM411 V 2.59 2016-09-29 - RCM Mobile v1.5.1)','Latitude','Longitude'], inplace=True, axis=1)\ndf = f\ndf.head(25)","metadata":{"id":"WdLw7lGTuw6v","outputId":"5a03dd8c-0456-4ddf-b8af-4a070ecf7d23","scrolled":true,"execution":{"iopub.status.busy":"2021-05-20T19:01:12.852819Z","iopub.execute_input":"2021-05-20T19:01:12.853212Z","iopub.status.idle":"2021-05-20T19:01:12.998245Z","shell.execute_reply.started":"2021-05-20T19:01:12.853176Z","shell.execute_reply":"2021-05-20T19:01:12.997146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df2 = df['Date']\nblankIndex=[''] * len(df2)\ndf2.index=blankIndex\ndata = str(df2).replace(\".\",\"-\")\npd.DataFrame([x.split(';') for x in data.split('\\n')])","metadata":{"execution":{"iopub.status.busy":"2021-05-20T19:02:18.119177Z","iopub.execute_input":"2021-05-20T19:02:18.119584Z","iopub.status.idle":"2021-05-20T19:02:18.601544Z","shell.execute_reply.started":"2021-05-20T19:02:18.119552Z","shell.execute_reply":"2021-05-20T19:02:18.600596Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df2 = df['Date']\ndata = str(df2).replace(\".\",\"-\")\npd.DataFrame([x for x in data.split('\\n')]).head(15)","metadata":{"execution":{"iopub.status.busy":"2021-05-20T19:06:45.106923Z","iopub.execute_input":"2021-05-20T19:06:45.107287Z","iopub.status.idle":"2021-05-20T19:06:45.120024Z","shell.execute_reply.started":"2021-05-20T19:06:45.107254Z","shell.execute_reply":"2021-05-20T19:06:45.119037Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df['Date']=pd.DataFrame([x.split(';') for x in data.split('\\n')])\ndf.head(25)","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:58:53.974017Z","iopub.execute_input":"2021-05-20T18:58:53.974635Z","iopub.status.idle":"2021-05-20T18:58:54.092135Z","shell.execute_reply.started":"2021-05-20T18:58:53.974582Z","shell.execute_reply":"2021-05-20T18:58:54.091125Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df.info()","metadata":{"id":"J_rm7nr4t3OQ","outputId":"f966af86-0e71-48e1-83e3-ee71a299de61","execution":{"iopub.status.busy":"2021-05-20T18:19:58.674401Z","iopub.execute_input":"2021-05-20T18:19:58.674689Z","iopub.status.idle":"2021-05-20T18:19:58.691746Z","shell.execute_reply.started":"2021-05-20T18:19:58.674662Z","shell.execute_reply":"2021-05-20T18:19:58.690784Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df.describe().transpose()","metadata":{"id":"YGpcqTZtt3RN","outputId":"a91c7312-3e77-48d8-89bc-7464c62216b5","execution":{"iopub.status.busy":"2021-05-20T18:19:58.692917Z","iopub.execute_input":"2021-05-20T18:19:58.693224Z","iopub.status.idle":"2021-05-20T18:19:59.577841Z","shell.execute_reply.started":"2021-05-20T18:19:58.693195Z","shell.execute_reply":"2021-05-20T18:19:59.576826Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df.isnull().sum()","metadata":{"id":"Q_za3I-vt3Tx","outputId":"35524285-1fba-4baa-ffc7-4ca532278874","execution":{"iopub.status.busy":"2021-05-20T18:19:59.57936Z","iopub.execute_input":"2021-05-20T18:19:59.57976Z","iopub.status.idle":"2021-05-20T18:20:00.031338Z","shell.execute_reply.started":"2021-05-20T18:19:59.57972Z","shell.execute_reply":"2021-05-20T18:20:00.030061Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Visualize our missing data\n\nmsno.bar(df, color=\"dodgerblue\")\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:00.034473Z","iopub.execute_input":"2021-05-20T18:20:00.034831Z","iopub.status.idle":"2021-05-20T18:20:01.48277Z","shell.execute_reply.started":"2021-05-20T18:20:00.0348Z","shell.execute_reply":"2021-05-20T18:20:01.481569Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df = df.dropna()\nmsno.bar(df, color=\"dodgerblue\")\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:01.48508Z","iopub.execute_input":"2021-05-20T18:20:01.48543Z","iopub.status.idle":"2021-05-20T18:20:03.635214Z","shell.execute_reply.started":"2021-05-20T18:20:01.4854Z","shell.execute_reply":"2021-05-20T18:20:03.633867Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df.drop_duplicates(subset=['Friction','Speed','Direction','Height'],keep='first',inplace=False,ignore_index=False)","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:03.636915Z","iopub.execute_input":"2021-05-20T18:20:03.637324Z","iopub.status.idle":"2021-05-20T18:20:04.218977Z","shell.execute_reply.started":"2021-05-20T18:20:03.637267Z","shell.execute_reply":"2021-05-20T18:20:04.217778Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:04.220659Z","iopub.execute_input":"2021-05-20T18:20:04.221091Z","iopub.status.idle":"2021-05-20T18:20:04.253035Z","shell.execute_reply.started":"2021-05-20T18:20:04.221043Z","shell.execute_reply":"2021-05-20T18:20:04.252158Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df.iloc[:,2:18]","metadata":{"id":"oEc2QPAWt3Y2","outputId":"6a7ffe9a-7ce3-4d79-9a9c-7a0fc1e18414","execution":{"iopub.status.busy":"2021-05-20T18:20:04.254265Z","iopub.execute_input":"2021-05-20T18:20:04.254557Z","iopub.status.idle":"2021-05-20T18:20:04.355806Z","shell.execute_reply.started":"2021-05-20T18:20:04.254531Z","shell.execute_reply":"2021-05-20T18:20:04.354775Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Accident_rates1 = [0.8, 0.55, 0.25, 0.2]\nAccident_rates = []\nfor x in df['Friction']:\n    if x < 0.15:\n        Accident_rates.append(0.8)\n    elif (x>=0.15 and x<=0.24):\n        Accident_rates.append(0.55)\n    elif(x>=0.25 and x<=0.34):\n        Accident_rates.append(0.25)\n    elif(x>=0.35 and x<=0.45):\n        Accident_rates.append(0.2)\n    else:\n        Accident_rates.append(0.05) #for preventive meaasures","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:04.357118Z","iopub.execute_input":"2021-05-20T18:20:04.357419Z","iopub.status.idle":"2021-05-20T18:20:05.547224Z","shell.execute_reply.started":"2021-05-20T18:20:04.357389Z","shell.execute_reply":"2021-05-20T18:20:05.546121Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df['Accident_rates_friction'] = Accident_rates\ndf.tail()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:05.548746Z","iopub.execute_input":"2021-05-20T18:20:05.549068Z","iopub.status.idle":"2021-05-20T18:20:05.963241Z","shell.execute_reply.started":"2021-05-20T18:20:05.549038Z","shell.execute_reply":"2021-05-20T18:20:05.962208Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Accident_rates_st = []\nfor x in df['State']:\n    if x == 1:\n        Accident_rates_st.append(0.12)#Dry\n    elif x == 2:\n        Accident_rates_st.append(0.16)#Moist\n    elif x == 3:\n        Accident_rates_st.append(0.16)#Wet\n    elif x == 4:\n        Accident_rates_st.append(0.53)#Icy\n    elif x == 5:\n        Accident_rates_st.append(0.3) #Snowy\n    elif x == 6:\n        Accident_rates_st.append(0.18)#Slushy","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:05.964681Z","iopub.execute_input":"2021-05-20T18:20:05.965003Z","iopub.status.idle":"2021-05-20T18:20:06.914787Z","shell.execute_reply.started":"2021-05-20T18:20:05.964966Z","shell.execute_reply":"2021-05-20T18:20:06.91363Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df['Accident_rates_state'] = Accident_rates_st","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:06.91606Z","iopub.execute_input":"2021-05-20T18:20:06.916349Z","iopub.status.idle":"2021-05-20T18:20:07.306075Z","shell.execute_reply.started":"2021-05-20T18:20:06.916321Z","shell.execute_reply":"2021-05-20T18:20:07.304778Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"month=[]\nfor x in df.iloc[:,0]:\n    month.append(x.split('.')[1])\n    #year,month,day=x.split('.')","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:07.307996Z","iopub.execute_input":"2021-05-20T18:20:07.308459Z","iopub.status.idle":"2021-05-20T18:20:08.506632Z","shell.execute_reply.started":"2021-05-20T18:20:07.308408Z","shell.execute_reply":"2021-05-20T18:20:08.505591Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df['Month']=month","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:08.50797Z","iopub.execute_input":"2021-05-20T18:20:08.508253Z","iopub.status.idle":"2021-05-20T18:20:08.682772Z","shell.execute_reply.started":"2021-05-20T18:20:08.508226Z","shell.execute_reply":"2021-05-20T18:20:08.681843Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df['Safety'] = df['Accident_rates_friction'] * df['Accident_rates_state']\ndf.head()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:08.684013Z","iopub.execute_input":"2021-05-20T18:20:08.684309Z","iopub.status.idle":"2021-05-20T18:20:08.77583Z","shell.execute_reply.started":"2021-05-20T18:20:08.684283Z","shell.execute_reply":"2021-05-20T18:20:08.774591Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"np.random.seed(0)\nimport seaborn as sns\nsns.set_theme()\nplt.figure(figsize= (15,10))\n_ = sns.heatmap(df.iloc[:,2:].corr(), annot= True, cmap= 'coolwarm')","metadata":{"id":"RHaxyETjt3WS","outputId":"bad2bf4f-153f-4a58-b78f-d4eb88b3ee24","execution":{"iopub.status.busy":"2021-05-20T18:20:08.777104Z","iopub.execute_input":"2021-05-20T18:20:08.7774Z","iopub.status.idle":"2021-05-20T18:20:11.46473Z","shell.execute_reply.started":"2021-05-20T18:20:08.777367Z","shell.execute_reply":"2021-05-20T18:20:11.463446Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"'''\ndf['Time(+01:00)'] = pd.to_datetime(df['Time(+01:00)'], format='%H:%M:%S').dt.hour\ndf.groupby(['Date','Time(+01:00)']).mean()\ndf = df.drop_duplicates()\ndf.head()\n'''","metadata":{"id":"2-pe5B0DND4Y","outputId":"3b7fd2d9-50e0-4d73-9030-ae52bc557384","execution":{"iopub.status.busy":"2021-05-20T18:20:11.466409Z","iopub.execute_input":"2021-05-20T18:20:11.467022Z","iopub.status.idle":"2021-05-20T18:20:11.473681Z","shell.execute_reply.started":"2021-05-20T18:20:11.466966Z","shell.execute_reply":"2021-05-20T18:20:11.472738Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# DISTRIBUTION OF THE DEPENDENT VARIABLE\n\nplt.figure(figsize=(20,5))\nsns.distplot(df['State'], bins=40, kde=True, color='red')\nplt.title('State Distribution')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:11.475369Z","iopub.execute_input":"2021-05-20T18:20:11.476097Z","iopub.status.idle":"2021-05-20T18:20:21.733676Z","shell.execute_reply.started":"2021-05-20T18:20:11.476053Z","shell.execute_reply":"2021-05-20T18:20:21.732929Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# States by different variables\n\nfig, ax = plt.subplots(2, 2, figsize= (10,10))\nax[0,0].scatter(df['Friction'], df['State'])\nax[0,0].set_title('State by friction')\nax[0,1].scatter(df['Ta'], df['State'])\nax[0,1].set_title('State by Ta')\nax[1,0].scatter(df['Height'], df['State'])\nax[1,0].set_title('State by Height')\nax[1,1].scatter(df['Speed'], df['State'])\nax[1,1].set_title('State by Speed')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:21.736407Z","iopub.execute_input":"2021-05-20T18:20:21.736891Z","iopub.status.idle":"2021-05-20T18:20:36.350379Z","shell.execute_reply.started":"2021-05-20T18:20:21.736841Z","shell.execute_reply":"2021-05-20T18:20:36.349707Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Decomposing Time-Series data into Trend and Seasonality","metadata":{}},{"cell_type":"code","source":"import math\nts= df.groupby(\"Date\")[\"State\"].mean().round(decimals=0)","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:36.351628Z","iopub.execute_input":"2021-05-20T18:20:36.352168Z","iopub.status.idle":"2021-05-20T18:20:36.546242Z","shell.execute_reply.started":"2021-05-20T18:20:36.35212Z","shell.execute_reply":"2021-05-20T18:20:36.545463Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(df.groupby(\"Date\")[\"State\"].mean())","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:30:52.121129Z","iopub.execute_input":"2021-05-20T18:30:52.121489Z","iopub.status.idle":"2021-05-20T18:30:52.33201Z","shell.execute_reply.started":"2021-05-20T18:30:52.121459Z","shell.execute_reply":"2021-05-20T18:30:52.330969Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Visualize Residuals, Seasonal, Trend, and level\n\nres = sm.tsa.seasonal_decompose(ts.values,period=7,model=\"multiplicative\")\nres.plot()\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:20:36.784322Z","iopub.execute_input":"2021-05-20T18:20:36.784609Z","iopub.status.idle":"2021-05-20T18:20:37.56672Z","shell.execute_reply.started":"2021-05-20T18:20:36.784582Z","shell.execute_reply":"2021-05-20T18:20:37.565762Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pd.date_range(\"2020-11-30\",\"2021-03-03\", freq=\"AS\").tolist()\nimport datetime\ndatetime.ts[].isoformat()","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:22:52.448087Z","iopub.execute_input":"2021-05-20T18:22:52.448443Z","iopub.status.idle":"2021-05-20T18:22:52.45591Z","shell.execute_reply.started":"2021-05-20T18:22:52.448413Z","shell.execute_reply":"2021-05-20T18:22:52.454869Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for i in range(64):\n    ts_year = ts[i:i+1]\n   # y=ts_year.values\n    #x=ts_year.index.day\n    print(ts_year.split())","metadata":{"execution":{"iopub.status.busy":"2021-05-20T18:38:00.92689Z","iopub.execute_input":"2021-05-20T18:38:00.927247Z","iopub.status.idle":"2021-05-20T18:38:00.955257Z","shell.execute_reply.started":"2021-05-20T18:38:00.927215Z","shell.execute_reply":"2021-05-20T18:38:00.953506Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Visualize the weekly sales by year\nfig = go.Figure()\nyears = pd.date_range(\"2020-11-30\",\"2021-03-03\", freq=\"H\").tolist() # range dates by year\nfor i in range(len(years)-1):\n    ts_year = ts[years[i]:years[i+1]]\n    fig.add_trace(\n        go.Scatter(\n            y=ts_year.values,\n            x=ts_year.index.week,\n            name=years[i].year,\n        ))\n\nfig.update_layout(\n    title=\"weekly sales by year\",\n    xaxis_title=\"weeks\",\n    yaxis_title=\"sales\",\n    legend_title=\"year\",\n    yaxis_tickprefix = '$',\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=18,\n        color=\"RebeccaPurple\"\n    ))\nfig.show()","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Target Variable: State \n### Features: Ta, Water, Tsurf, Speed, Direction","metadata":{}},{"cell_type":"code","source":"from sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.model_selection import train_test_split, GridSearchCV\nfrom sklearn.metrics import classification_report, confusion_matrix\n\nX = df.loc[:,['Ta','Water','Tsurf','Speed','Direction']]\ny = df['State']","metadata":{"id":"ixuc-hcPICaQ","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 42)\nknn = KNeighborsClassifier(n_neighbors = 6)\nknn.fit(X_train, y_train)\nprint(knn.score(X_test, y_test))","metadata":{"id":"d-52KwXtLhqq","outputId":"712f8980-924b-4448-d3a1-6796dd141dea","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 42)\n\nlogreg = LogisticRegression()\n\nc_space = np.logspace(-1,1,2)\nparam_grid = {'C': c_space}\nlogreg_cv = GridSearchCV(logreg, param_grid, cv=3)\n\nlogreg_cv.fit(X_train, y_train)\n\ny_pred = logreg_cv.predict(X_test)\n\nprint(confusion_matrix(y_test, y_pred))\nprint(classification_report(y_test, y_pred))\n\nprint(\"Tuned Logistic Regression Parameters: {}\".format(logreg_cv.best_params_)) \nprint(\"Best score is {}\".format(logreg_cv.best_score_))","metadata":{"id":"Vyvs9z5nLzyq","outputId":"5a311cd7-93c4-47b2-95ac-fb785abdca11","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.preprocessing import scale\nX_scaled = scale(X)\nprint(\"Mean of Scaled features {}\".format(np.mean(X_scaled)))\nprint(\"Standard deviation of Scaled features {}\".format(np.std(X_scaled)))","metadata":{"id":"7bfK1DyqpDkX","outputId":"6712ee0d-234d-4019-a668-c2e594d16786","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import cross_val_score\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.pipeline import Pipeline\n\nsteps = [('scaler', StandardScaler()),\n        ('knn', KNeighborsClassifier())]\n\npipeline = Pipeline(steps)\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n\n# Fit the pipeline to the training set: knn_scaled\nknn_scaled = pipeline.fit(X_train, y_train)\n\n# Instantiate and fit a k-NN classifier to the unscaled data\nknn_unscaled = KNeighborsClassifier().fit(X_train, y_train)\n\n# Compute and print metrics\nprint('Accuracy with Scaling: {}'.format(knn_scaled.score(X_test, y_test)))\nprint('Accuracy without Scaling: {}'.format(knn_unscaled.score(X_test, y_test)))","metadata":{"id":"omVm3sjAOb6U","outputId":"2273995b-1e96-47a4-9a22-159381f0b4d0","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.svm import SVC\nsteps = [('scaler', StandardScaler()),\n         ('SVM', SVC())]\n\npipeline = Pipeline(steps)\n\n# Specify the hyperparameter space\nparameters = {'SVM__C':[1, 10, 100],\n              'SVM__gamma':[0.1, 0.01]}\n\n# Create train and test sets\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=21)\n\n# Instantiate the GridSearchCV object: cv\ncv = GridSearchCV(pipeline, param_grid=parameters)\n\n# Fit to the training set\ncv.fit(X_train, y_train)\n\n# Predict the labels of the test set: y_pred\ny_pred = cv.predict(X_test)\n\n# Compute and print metrics\nprint(\"Accuracy: {}\".format(cv.score(X_test, y_test)))\nprint(classification_report(y_test, y_pred))\nprint(\"Tuned Model Parameters: {}\".format(cv.best_params_))","metadata":{"id":"d6d7_DZ7Xvnb","outputId":"ff31cf26-99ed-4f9d-be36-cd387420fb61","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Target Variable: State \n### Features: Ta, Water, Tsurf, Speed, Direction, Height, Friction","metadata":{}},{"cell_type":"code","source":"from sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.model_selection import train_test_split, GridSearchCV\nfrom sklearn.metrics import classification_report, confusion_matrix\n\nX = df.loc[:,['Ta','Water','Tsurf','Speed','Direction','Height','Friction']]\ny = df['State']","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import cross_val_score\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.pipeline import Pipeline\n\nsteps = [('scaler', StandardScaler()),\n        ('knn', KNeighborsClassifier())]\n\npipeline = Pipeline(steps)\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n\n# Fit the pipeline to the training set: knn_scaled\nknn_scaled = pipeline.fit(X_train, y_train)\n\n# Instantiate and fit a k-NN classifier to the unscaled data\nknn_unscaled = KNeighborsClassifier().fit(X_train, y_train)\n\n# Compute and print metrics\nprint('Accuracy with Scaling: {}'.format(knn_scaled.score(X_test, y_test)))\nprint('Accuracy without Scaling: {}'.format(knn_unscaled.score(X_test, y_test)))","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.svm import SVC\nsteps = [('scaler', StandardScaler()),\n         ('SVM', SVC())]\n\npipeline = Pipeline(steps)\n\n# Specify the hyperparameter space\nparameters = {'SVM__C':[100],\n              'SVM__gamma':[0.1]}\n\n# Create train and test sets\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=21)\n\n# Instantiate the GridSearchCV object: cv\ncv = GridSearchCV(pipeline, param_grid=parameters)\n\n# Fit to the training set\ncv.fit(X_train, y_train)\n\n# Predict the labels of the test set: y_pred\ny_pred = cv.predict(X_test)\n\n# Compute and print metrics\nprint(\"Accuracy: {}\".format(cv.score(X_test, y_test)))\nprint(classification_report(y_test, y_pred))\nprint(\"Tuned Model Parameters: {}\".format(cv.best_params_))","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"### Target Variable: State \n### Features: Ta, Water, Tsurf, Speed, Direction, Height, Friction\nAccuracy with Scaling: 0.93762372198151\nAccuracy without Scaling: 0.6891010419981559","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Target Variable: Safety = Accident rate in case of state * Accident rate in case of friction coefficient","metadata":{"id":"Bc_fgOm6Z_7z"}},{"cell_type":"code","source":"from sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.model_selection import train_test_split, GridSearchCV\nfrom sklearn.metrics import classification_report, confusion_matrix\n\nX = df.loc[:,['Ta','Water','Tsurf','Speed','Direction','Height']]\ny = df['Safety']","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.linear_model import ElasticNet\nsteps = [('scaler', StandardScaler()),\n         ('elasticnet', ElasticNet())]\n\n# Create the pipeline: pipeline \npipeline = Pipeline(steps)\n\n# Specify the hyperparameter space\nparameters = {'elasticnet__l1_ratio':np.linspace(0,1,30)}\n\n# Create train and test sets\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, random_state=42)\n\n# Create the GridSearchCV object: gm_cv\ngm_cv = GridSearchCV(pipeline, param_grid=parameters)\n\n# Fit to the training set\ngm_cv.fit(X_train, y_train)\n\n# Compute and print the metrics\nr2 = gm_cv.score(X_test, y_test)\nprint(\"Tuned ElasticNet Alpha: {}\".format(gm_cv.best_params_))\nprint(\"Tuned ElasticNet R squared: {}\".format(r2))","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}