-
Notifications
You must be signed in to change notification settings - Fork 0
/
2013-02-22-Decision-Trees.html
763 lines (602 loc) · 20.7 KB
/
2013-02-22-Decision-Trees.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
<head>
<title>2013-02-22-Decision-Trees</title>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8"/>
<meta name="title" content="2013-02-22-Decision-Trees"/>
<meta name="generator" content="Org-mode"/>
<meta name="generated" content="2013-02-22 08:41:02 PST"/>
<meta name="author" content="Jim Blomo"/>
<meta name="description" content=""/>
<meta name="keywords" content=""/>
<link rel="stylesheet" type="text/css" href="production/common.css" />
<link rel="stylesheet" type="text/css" href="production/screen.css" media="screen" />
<link rel="stylesheet" type="text/css" href="production/projection.css" media="projection" />
<link rel="stylesheet" type="text/css" href="production/color-blue.css" media="projection" />
<link rel="stylesheet" type="text/css" href="production/presenter.css" media="presenter" />
<link href='http://fonts.googleapis.com/css?family=Lobster+Two:700|Yanone+Kaffeesatz:700|Open+Sans' rel='stylesheet' type='text/css'>
</head>
<body>
<div id="preamble">
</div>
<div id="content">
<h1 class="title">2013-02-22-Decision-Trees</h1>
<div id="table-of-contents">
<h2>Table of Contents</h2>
<div id="text-table-of-contents">
<ul>
<li><a href="#sec-1">1 Classification: Decision Trees</a></li>
<li><a href="#sec-2">2 Types of Models</a>
<ul>
<li><a href="#sec-2-1">2.1 Details</a></li>
</ul>
</li>
<li><a href="#sec-3">3 Process</a>
<ul>
<li><a href="#sec-3-1">3.1 Steps</a></li>
</ul>
</li>
<li><a href="#sec-4">4 Learning</a>
<ul>
<li><a href="#sec-4-1">4.1 Example</a></li>
</ul>
</li>
<li><a href="#sec-5">5 Classification</a>
<ul>
<li><a href="#sec-5-1">5.1 Possibilities</a></li>
</ul>
</li>
<li><a href="#sec-6">6 Machine Learning</a>
<ul>
<li><a href="#sec-6-1">6.1 Which is this?</a></li>
</ul>
</li>
<li><a href="#sec-7">7 Confusion Matrix</a>
<ul>
<li><a href="#sec-7-1">7.1 Basis for Evaluation</a></li>
</ul>
</li>
<li><a href="#sec-8">8 Recall & Precision</a>
<ul>
<li><a href="#sec-8-1">8.1 Trade-off</a></li>
</ul>
</li>
<li><a href="#sec-9">9 Example: Search</a>
<ul>
<li><a href="#sec-9-1">9.1 Searching Yelp</a></li>
</ul>
</li>
<li><a href="#sec-10">10 Decision Trees</a>
<ul>
<li><a href="#sec-10-1">10.1 Rules expressed trees</a></li>
</ul>
</li>
<li><a href="#sec-11">11 Build a Tree</a>
<ul>
<li><a href="#sec-11-1">11.1 Directions</a></li>
</ul>
</li>
<li><a href="#sec-12">12 Build a Tree</a>
<ul>
<li><a href="#sec-12-1">12.1 Next challenge</a></li>
<li><a href="#sec-12-2">12.2 Build a Tree</a>
<ul>
<li><a href="#sec-12-2-1">12.2.1 Result</a></li>
</ul></li>
</ul>
</li>
<li><a href="#sec-13">13 Decision Tree Induction</a>
<ul>
<li><a href="#sec-13-1">13.1 Recursive</a></li>
</ul>
</li>
<li><a href="#sec-14">14 Information Gain</a>
<ul>
<li><a href="#sec-14-1">14.1 Information</a></li>
</ul>
</li>
<li><a href="#sec-15">15 Gini Index</a></li>
<li><a href="#sec-16">16 Splitting</a>
<ul>
<li><a href="#sec-16-1">16.1 Different</a></li>
</ul>
</li>
<li><a href="#sec-17">17 Decision Tree Advantages</a>
<ul>
<li><a href="#sec-17-1">17.1 Trees</a></li>
</ul>
</li>
<li><a href="#sec-18">18 <b>Break</b></a></li>
</ul>
</div>
</div>
<div id="outline-container-1" class="outline-2">
<h2 id="sec-1"><span class="section-number-2">1</span> Classification: Decision Trees <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-1">
</div>
</div>
<div id="outline-container-2" class="outline-2">
<h2 id="sec-2"><span class="section-number-2">2</span> Types of Models <span class="tag"><span class="slide">slide</span> <span class="animate">animate</span></span></h2>
<div class="outline-text-2" id="text-2">
<ul>
<li>Classifiers
</li>
<li>Regressions
</li>
<li>Clustering
</li>
<li>Outlier
</li>
</ul>
</div>
<div id="outline-container-2-1" class="outline-3">
<h3 id="sec-2-1"><span class="section-number-3">2.1</span> Details <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-2-1">
<dl>
<dt>Classifiers</dt><dd>describes and distinguishes cases. Yelp may want to find a
category for a business based on the reviews and business description
</dd>
<dt>Regressions</dt><dd>Predict a continuous value. Eg. predict a home's selling
price given sq footage, # of bedrooms
</dd>
<dt>Clustering</dt><dd>find "natural" groups of data <b>without labels</b>
</dd>
<dt>Outlier</dt><dd>find anomalous transactions, eg. finding fraud for credit cards
</dd>
</dl>
</div>
</div>
</div>
<div id="outline-container-3" class="outline-2">
<h2 id="sec-3"><span class="section-number-2">3</span> Process <span class="tag"><span class="slide">slide</span> <span class="animate">animate</span></span></h2>
<div class="outline-text-2" id="text-3">
<ul>
<li>Training Set
</li>
<li>Learning
</li>
<li>Model / Classifier
</li>
<li>Testing Set
</li>
<li>Verification / Accuracy
</li>
<li>New Data
</li>
<li>Classification
</li>
</ul>
</div>
<div id="outline-container-3-1" class="outline-3">
<h3 id="sec-3-1"><span class="section-number-3">3.1</span> Steps <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-3-1">
<dl>
<dt>Process</dt><dd>to be able to classify data
</dd>
<dt>Training Set</dt><dd>Cleaned, preprocessed data that has labels. What are
labels?
</dd>
<dt>Learning</dt><dd>Feed the training set to an algorithm. Algorithm associates
some of the features with the labels and generates a model.
</dd>
<dt>Model / Classifier</dt><dd>Process or formula used to predict the label (class) given inputs
(data record)
</dd>
<dt>Testing Set</dt><dd>Data <b>not in training set</b>, with labels. Run through model
to see how the model compares with the real labels.
</dd>
<dt>Verification / Accuracy</dt><dd>Given the matches / mismatches in the testing
set, how can we measure how well the model reflects reality?
</dd>
<dt>Unseen Data</dt><dd>Finally, we're ready to start using our model / classifier to
label new, real, unknown data! So clean and pre-process it the same way.
</dd>
<dt>Classification</dt><dd>Feed the unknown data and get out results!
</dd>
</dl>
</div>
</div>
</div>
<div id="outline-container-4" class="outline-2">
<h2 id="sec-4"><span class="section-number-2">4</span> Learning <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-4">
<p> <img src="img/model.png" alt="img/model.png" />
</p>
</div>
<div id="outline-container-4-1" class="outline-3">
<h3 id="sec-4-1"><span class="section-number-3">4.1</span> Example <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-4-1">
<ul>
<li>We have training data. What are these column types?
</li>
<li>Feed it into a classification algorithm
</li>
<li>In the case it is generating Rules.
</li>
<li>Models can be as simple as this: just a set of rules to follow. We'll see
how we can extend this idea
</li>
<li>The learning step generates a model: these rules
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-5" class="outline-2">
<h2 id="sec-5"><span class="section-number-2">5</span> Classification <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-5">
<p> <img src="img/classifying.png" alt="img/classifying.png" />
</p>
</div>
<div id="outline-container-5-1" class="outline-3">
<h3 id="sec-5-1"><span class="section-number-3">5.1</span> Possibilities <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-5-1">
<ul>
<li>Now that we have the model / classifier, we can do two things
</li>
<li>1: Use testing data <b>different</b> from training data
</li>
<li>compare the classifier guesses with reality
</li>
<li>2: Use the classifier on unknown data
</li>
<li>Why not just jump into classifying unknown data? Why have a test step?
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-6" class="outline-2">
<h2 id="sec-6"><span class="section-number-2">6</span> Machine Learning <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-6">
<dl>
<dt>Supervised</dt><dd>Given data with a label, predict data without a
label
</dd>
<dt>Unsupervised</dt><dd>Given data without labels, group "similar" items
together
</dd>
<dt>Semi-supervised</dt><dd>Mix of the above: eg. unsupervised to find groups,
supervised to label and distinguish borderline cases
</dd>
<dt>Active</dt><dd>Starting with unlabeled data, select the most helpful cases for a
human to label
</dd>
</dl>
</div>
<div id="outline-container-6-1" class="outline-3">
<h3 id="sec-6-1"><span class="section-number-3">6.1</span> Which is this? <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-6-1">
<ul>
<li>In the example above, what type of learning?
</li>
<li>Supervised: we have labels, we want to guess unlabeled data
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-7" class="outline-2">
<h2 id="sec-7"><span class="section-number-2">7</span> Confusion Matrix <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-7">
<ul>
<li>What are the ways that classification can be wrong?
</li>
</ul>
<table border="2" cellspacing="0" cellpadding="6" rules="groups" frame="hsides">
<caption></caption>
<colgroup><col class="left" /><col class="left" /><col class="left" />
</colgroup>
<tbody>
<tr><td class="left"></td><td class="left">Predict: Positive</td><td class="left">Predict: Negative</td></tr>
<tr><td class="left">Actual: Positive</td><td class="left">True Positive</td><td class="left">False Negative</td></tr>
<tr><td class="left">Actual: Negative</td><td class="left">False Negative</td><td class="left">True Negative</td></tr>
</tbody>
</table>
</div>
<div id="outline-container-7-1" class="outline-3">
<h3 id="sec-7-1"><span class="section-number-3">7.1</span> Basis for Evaluation <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-7-1">
<ul>
<li>Most methods of evaluating results start with the confusion matrix
</li>
<li>Figuring out what different ways you were right or wrong
</li>
<li>Then using different formulas to emphasize the things you care about
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-8" class="outline-2">
<h2 id="sec-8"><span class="section-number-2">8</span> Recall & Precision <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-8">
<ul>
<li>Recall: <code>TP / P</code>
</li>
<li>Precision: <code>TP / (TP + FP)</code>
</li>
<li>Sometimes these are in tension; other measurements balance them
</li>
</ul>
</div>
<div id="outline-container-8-1" class="outline-3">
<h3 id="sec-8-1"><span class="section-number-3">8.1</span> Trade-off <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-8-1">
<ul>
<li>Classic trade-off in search
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-9" class="outline-2">
<h2 id="sec-9"><span class="section-number-2">9</span> Example: Search <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-9">
<p> <img src="img/burrito-search.png" alt="img/burrito-search.png" />
</p>
</div>
<div id="outline-container-9-1" class="outline-3">
<h3 id="sec-9-1"><span class="section-number-3">9.1</span> Searching Yelp <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-9-1">
<ul>
<li>Searched yelp for a burrito in the Mission
</li>
<li>How good are these search results?
</li>
<li>Let's say we knew this first result was great, and <b>only</b> returned it
</li>
<li>What would our precision be?
</li>
<li>What would the recall be?
</li>
<li>How could we improve recall?
</li>
<li>How can we guarantee 100% recall?
</li>
<li>What will that do to the precision?
</li>
<li>Understand ways of combining these measurements in the book
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-10" class="outline-2">
<h2 id="sec-10"><span class="section-number-2">10</span> Decision Trees <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-10">
<ul>
<li>Rules formulated as a tree of decisions
</li>
<li>Choose Your Own Adventure for machine learning
</li>
<li>So how do we build the trees?
</li>
</ul>
</div>
<div id="outline-container-10-1" class="outline-3">
<h3 id="sec-10-1"><span class="section-number-3">10.1</span> Rules expressed trees <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-10-1">
<ul>
<li>At each node in the tree, pose a question
</li>
<li>Take a branch depending on your answer
</li>
<li>Leaf nodes are labels
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-11" class="outline-2">
<h2 id="sec-11"><span class="section-number-2">11</span> Build a Tree <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-11">
<p> <img src="img/model.png" alt="img/model.png" />
</p>
</div>
<div id="outline-container-11-1" class="outline-3">
<h3 id="sec-11-1"><span class="section-number-3">11.1</span> Directions <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-11-1">
<ul>
<li>First node question: is rank=professor?
</li>
<li>If True, what's the label?
</li>
<li>If False, we go to another node
</li>
<li>Second node question: is years > 6?
</li>
<li>If True what's the label?
</li>
<li>If False, what's the label?
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-12" class="outline-2">
<h2 id="sec-12"><span class="section-number-2">12</span> Build a Tree <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-12">
<p> <img src="img/tree-dataset.png" alt="img/tree-dataset.png" />
</p>
</div>
<div id="outline-container-12-1" class="outline-3">
<h3 id="sec-12-1"><span class="section-number-3">12.1</span> Next challenge <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-12-1">
<ul>
<li>How to go from a data set like this
</li>
</ul>
</div>
</div>
<div id="outline-container-12-2" class="outline-3">
<h3 id="sec-12-2"><span class="section-number-3">12.2</span> Build a Tree <span class="tag"><span class="slide">slide</span></span></h3>
<div class="outline-text-3" id="text-12-2">
<p> <img src="img/tree.png" alt="img/tree.png" />
</p>
</div>
<div id="outline-container-12-2-1" class="outline-4">
<h4 id="sec-12-2-1"><span class="section-number-4">12.2.1</span> Result <span class="tag"><span class="notes">notes</span></span></h4>
<div class="outline-text-4" id="text-12-2-1">
<ul>
<li>To a tree like this?
</li>
</ul>
</div>
</div>
</div>
</div>
<div id="outline-container-13" class="outline-2">
<h2 id="sec-13"><span class="section-number-2">13</span> Decision Tree Induction <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-13">
<ul>
<li>Start with all the data
</li>
<li>Choose the "best" way to divide it up based on one attribute
</li>
<li>Make a node that asks a question to split the data
</li>
<li>Choose new "best" way to divide based on remaining attributes
</li>
<li>Stop: no attributes left, all records are the same class
</li>
</ul>
</div>
<div id="outline-container-13-1" class="outline-3">
<h3 id="sec-13-1"><span class="section-number-3">13.1</span> Recursive <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-13-1">
<ul>
<li>Look at all the attributes. What's the best way to split up the data?
</li>
<li>We'll look at way to mathematically evaluate splits
</li>
<li>Now recursively do the same
</li>
<li>If you've split on all the attributes, but still have a mix, use a majority
rule
</li>
<li>If all the records are the same class, you don't have to keep spitting:
your answer is right there!
</li>
<li>For continuous data, must bucket it so you can have a discrete number of
answers
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-14" class="outline-2">
<h2 id="sec-14"><span class="section-number-2">14</span> Information Gain <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-14">
<ul>
<li>Comparison of how mixed results are before and after splitting
</li>
<li>Entropy measurement of "mixed"
</li>
<li>Two pure data sets have less entropy on average than one mixed
</li>
</ul>
</div>
<div id="outline-container-14-1" class="outline-3">
<h3 id="sec-14-1"><span class="section-number-3">14.1</span> Information <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-14-1">
<ul>
<li>Book will go into detail about how to think about entropy
</li>
<li>General idea: how difficult would it be to memorize the data sets?
</li>
<li>Easy if pure: all class A
</li>
<li>Still fairly easy if 2 pure sets: 1 is class A, other is class B
</li>
<li>Now more difficult if they are mixed: first 2 records are A, then one B,
then another A
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-15" class="outline-2">
<h2 id="sec-15"><span class="section-number-2">15</span> Gini Index <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-15">
<pre class="src src-python">Gini(D) = 1 - <span style="color: #7a378b;">sum</span>(frac**2 <span style="color: #a020f0;">for</span> frac <span style="color: #a020f0;">in</span> classes)
</pre>
<p>
Sum of the squares of the fraction of items in each class
</p>
</div>
</div>
<div id="outline-container-16" class="outline-2">
<h2 id="sec-16"><span class="section-number-2">16</span> Splitting <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-16">
<ul>
<li>Discrete values can split per value
</li>
<li>Or discrete values binary split into subsets
</li>
<li>Continuous values can split on range (usually 2)
</li>
</ul>
</div>
<div id="outline-container-16-1" class="outline-3">
<h3 id="sec-16-1"><span class="section-number-3">16.1</span> Different <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-16-1">
<ul>
<li>If you'd like a binary tree (useful for some algorithms), can split on
subsets
</li>
<li>Can't split 400 different ways on continuous values… what about values
that haven't been seen before?
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-17" class="outline-2">
<h2 id="sec-17"><span class="section-number-2">17</span> Decision Tree Advantages <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-17">
<ul>
<li>Models easy to understand and visualize
</li>
<li>Can be faster to construct
</li>
<li>Can encode tree in declarative languages (SQL)
</li>
<li>Robust: outliers generally fit in with normal data
</li>
</ul>
</div>
<div id="outline-container-17-1" class="outline-3">
<h3 id="sec-17-1"><span class="section-number-3">17.1</span> Trees <span class="tag"><span class="notes">notes</span></span></h3>
<div class="outline-text-3" id="text-17-1">
<ul>
<li>Its a tree! Easy to draw
</li>
<li>Greedy algorithm means you're only go over the data so many times
</li>
<li>Models can translate into database statements
</li>
<li>Outliers don't have a numeric pull on the data (similar to difference
between median and mean)
</li>
</ul>
</div>
</div>
</div>
<div id="outline-container-18" class="outline-2">
<h2 id="sec-18"><span class="section-number-2">18</span> <b>Break</b> <span class="tag"><span class="slide">slide</span></span></h2>
<div class="outline-text-2" id="text-18">
<script type="text/javascript" src="production/org-html-slideshow.js"></script>
</div>
</div>
</div>
<div id="postamble">
<p class="date">Date: 2013-02-22 08:41:02 PST</p>
<p class="author">Author: Jim Blomo</p>
<p class="creator">Org version 7.8.02 with Emacs version 23</p>
<a href="http://validator.w3.org/check?uri=referer">Validate XHTML 1.0</a>
</div>
</body>
</html>