-
Notifications
You must be signed in to change notification settings - Fork 9
/
Loan_behavior_prediction.ipynb
776 lines (776 loc) · 245 KB
/
Loan_behavior_prediction.ipynb
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
764
765
766
767
768
769
770
771
772
773
774
775
776
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"<a class=\"anchor\" id=\"top\"></a>\n",
"### Predicting bank loan behavior by analyzing bank transactional data with a random forest classifier\n",
"Andy Ganse, andy@ganse.org, http://research.ganse.org\n",
"\n",
"The title says it all -- let's implement a random forest classifier from [Scikit-learn](http://scikit-learn.org/stable) to see how well we can predict whether a bank client will have good loan behavior (meaning they won't default or become delinquent) if given a new loan. In the process we'll review the results validations for this kind of model, and learn about some limitations of Scikit-learn's models. \n",
"\n",
"Turns out publicly-available bank transaction data is really hard to come by, American or otherwise. For use below I've found a foreign dataset from the Czech Republic in 1993-1998 - it was a shared dataset for a 1999 European machine learning conference challenge. This was the PKDD99 Challenge, from the \"European Conferences on Machine Learning and European Conferences on Principles and Practice of Knowledge Discovery in Databases ECML/PKDD Discovery Challenges 1999 - 2005 : A Collaborative Effort in Knowledge Discovery from Databases\" (whew!). The dataset is still available on the old conference website (http://sorry.vse.cz/~berka/challenge/PAST) -- scroll down to \"PKDD'99 Challenge\" on its bottom left. The dataset description document describing the fields and format is at http://sorry.vse.cz/~berka/challenge/pkdd1999/berka.htm \n",
"\n",
"These data are anonymized bank transactions covering six years of bank transactions for 5369 bank clients with 4500 accounts (ie some of those clients are coupled on some of the accounts, as for example husbands and wifes). The currency, the koruna (crown) or CZK, varied in the 25-30/USD range in the 1990s during the dataset. Exchange rate data and plotted histories are available at https://tradingeconomics.com/czech-republic/currency. The dataset is not in English, has some funny encoding (eg combining birth-date and gender), and is contained in csv files, so I wrote a simple **module that translates and reorganizes the PKDD99 data into a set of Pandas dataframes all in English; these functions may be of interest to others for convenience in other work -- they're in [`pkd99_bank_data.py`](pkd99_bank_data.py)**.\n",
"\n",
"Some questions to explore below in our modeling of this dataset:\n",
"* what are the most important features in the data that are likely to determine poor loan behavior (default and delinquincy)?\n",
"* how accurately can we predict the loan behavior given these features?\n",
"* is the data's response variable balanaced in terms of frequency of positive and negative occurances, i.e. good loans and bad loans? Presumably not (hopefully not for the bank's sake!) and in that case does resampling help?\n",
"\n",
"\n",
"#### Contents: [Getting the data](#getting) | [Creating the features dataframe](#createfeatures) | [Classification](#classification) | [Next steps](#next) \n",
"<HR>"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"# the usual iPython boilerplate...\n",
"%matplotlib inline\n",
"from IPython.display import display\n",
"import matplotlib.pyplot as plt\n",
"import pandas as pd\n",
"import datetime\n",
"pd.set_option('display.width', 1000)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Getting the data <a class=\"anchor\" id=\"getting\"></a><small>(return to [top](#top))</small>"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Download and unzip the [dataset file](http://sorry.vse.cz/~berka/challenge/pkdd1999/data_berka.zip) from the website mentioned above. The zip file contains the CSV files listed below (with extension `.asc` for \"ASCII\" text). These `.asc` files must be placed in the same directory as this `.ipynb` notebook file before running it:"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"-rwxr-xr-x@ 1 aganse staff 155356 Mar 15 1999 \u001b[31maccount.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 31588 Mar 15 1999 \u001b[31mcard.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 94820 Mar 15 1999 \u001b[31mclient.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 129716 Mar 15 1999 \u001b[31mdisp.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 6574 Mar 15 1999 \u001b[31mdistrict.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 27037 Mar 15 1999 \u001b[31mloan.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 273800 Mar 15 1999 \u001b[31morder.asc\u001b[m\u001b[m\r\n",
"-rwxr-xr-x@ 1 aganse staff 69406578 Mar 15 1999 \u001b[31mtrans.asc\u001b[m\u001b[m\r\n"
]
}
],
"source": [
"# checking file sizes, eg note the transactions take approx 70MB\n",
"!ls -l *.asc"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Then we use this reader function I wrote to load up and translate and organize the data from those files into some Pandas dataframes:"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"import pkdd99_bank_data as pkdd99\n",
"account,card,client,disp,district,loan,order,trans = pkdd99.get_bank_data()\n",
"# note pkdd99.get_bank_data() took ~5min to run on my MBP2015/i7/16GB"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"These dataframes have metadata of `df.name`, `df.description`, `df.notes` which persist thru the functions in `PKDD99_bank_data` and this notebook, but DO NOT persist thru pickling (ie saving to file and reloading). This metadata persistance issue has been a recurring discussion in the Python dev community. I just found this metadata a convenient way to store and reread details about the data, but these details are found in the data description document - so as long as you don't mind losing this internal form of documentation then it's really not a big deal to pickle away. My own main motivation for saving files was that get_bank_data() took ~5min to run on my MBP/16GB and I didn't want to keep waiting for it while testing.\n",
"\n",
"Again the original [dataset description document](http://sorry.vse.cz/~berka/challenge/pkdd1999/berka.htm) describes the fields and contents, but I've translated and cleaned up the fields and contents in the Pandas dataframes, and their `table_summary()` function lets us see descriptions and the structure of these eight dataframes via top three records. Think of them as database tables with primary and foreign keys that link the contents across the tables:"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Client : length 5369 : each record describes characteristics of a client\n",
"(one client can have one or more accounts)\n",
" client_id date_birth district_id gender\n",
"0 1 1970-12-13 18 F\n",
"1 2 1945-02-04 1 M\n",
"2 3 1940-10-09 1 F\n",
" \n",
"Loan : length 682 : each record describes a loan granted for a given account\n",
"(one account may have zero or one loan. see loan.codes for ABCD status definitions)\n",
" loan_id account_id date amount duration payments status\n",
"0 5314 1787 1993-07-05 96396 12 8033.0 B\n",
"1 5316 1801 1993-07-11 165960 36 4610.0 A\n",
"2 6863 9188 1993-07-28 127080 60 2118.0 A\n",
" \n",
"Loan status codes:\n",
"A = contract finished, no problems\n",
"B = contract finished, loan not payed\n",
"C = running contract, OK so far\n",
"D = running contract, client in debt\n"
]
}
],
"source": [
"# Just outputting a few for sake of space...\n",
"#PKDD99.table_summary(account)\n",
"#PKDD99.table_summary(card)\n",
"#PKDD99.table_summary(disp)\n",
"#PKDD99.table_summary(district)\n",
"#PKDD99.table_summary(order)\n",
"#PKDD99.table_summary(trans)\n",
"PKDD99.table_summary(client)\n",
"PKDD99.table_summary(loan)\n",
"print(loan.codes)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here's the structure of those dataframe relationships:\n",
"<img src=\"structure.png\" width=\"60%\">"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Creating the features dataframe <a class=\"anchor\" id=\"createfeatures\"></a><small>(return to [top](#top))</small>\n",
"After having explored the dataset a bit I'm ready to create a features/response dataframe keyed on clients, to use in modeling loan behavior. Honestly I don't yet know whether it's more informative to key on each client vs on each account -- noting that one client can have more than one account, and one account can have more than one client -- in this work I'll just key on each client. In this preliminary analysis I'll use six of the eight data tables.\n",
"\n",
"First we do SQL-style merging of the data from the different dataframes to create a new client-features dataframe, and check the number of resulting records against what we expect given the lengths of the original tables/dataframes:"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"5369 total feature records, ie one for each client\n",
"827 feature records with a loan; some accts repeated due to multiple clients on same acct\n",
"682 feature records with a loan and unique account_id\n"
]
}
],
"source": [
"# Do SQL-style merge of data from different dataframes to create features dataframe for classification\n",
"features = pd.merge(account,disp,on='account_id',how='outer')\n",
"features = pd.merge(features,loan,on='account_id',how='left',suffixes=('_acct','_loan')) # sufficies for date\n",
"features = pd.merge(features,client,on='client_id',how='outer',suffixes=('_bank','_client')) # suffices for district_id\n",
"features = pd.merge(features,card,on='disp_id',how='outer',suffixes=('_disp','_card')) # suffices for type\n",
"features.rename(columns = {'date':'date_card'}, inplace=True) # for clarity\n",
"print(len(features),'total feature records, ie one for each client') # should be 5369\n",
"features = features[pd.notnull(features['loan_id'])]\n",
"print(len(features),'feature records with a loan; some accts repeated due to multiple clients on same acct') # should be 827\n",
"print(len(features['account_id'].unique()),'feature records with a loan and unique account_id') # should be 682"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next we compute features for each client from the transactions data - let's do min/max/mean of the account balance for the cumulative M months before the loan date. So `min1`, `max1`, and `mean1` are for the month before the loan; `min2`, `max2`, and `mean2` cover both of the two months before the loan (not just the second month before the loan); `min3`, `max3`, and `mean3` cover the three months before the loan. And so on."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"# Append columns of minbalM,maxbalM,meanbalM for the cumulative M months before loan start date.\n",
"# (note: technically it's M*30days rather than M months due to varying days/month...)\n",
"def addcols_monbalstats(features,trans,M):\n",
" trans_acctdate = pd.merge(trans,features[['account_id','date_loan']],on='account_id',how='inner')\n",
" # drop transactions that took place after the loan date:\n",
" trans_acctdate['datediff'] = trans_acctdate['date_loan'].subtract(trans_acctdate['date']) # 'date' = trans date\n",
" trans_acctdate.drop(trans_acctdate[trans_acctdate.datediff < datetime.timedelta(0)].index, inplace=True)\n",
" # reduce to transactions with M*30days of the loan date (ie datediff < M*30):\n",
" trans_acctdate.drop(trans_acctdate[trans_acctdate.datediff > datetime.timedelta(M*30)].index, inplace=True)\n",
" #monbalstats = trans_acctdate.groupby('account_id')['balance'].agg(['min','max','mean','count']).reset_index()\n",
" monbalstats = trans_acctdate.groupby('account_id')['balance'].agg(['min','max','mean']).reset_index()\n",
" #monbalstats.rename(columns = {'min':'min'+str(M),'max':'max'+str(M),'mean':'mean'+str(M),'count':'N'+str(M)}, inplace=True)\n",
" monbalstats.rename(columns = {'min':'min'+str(M),'max':'max'+str(M),'mean':'mean'+str(M)}, inplace=True)\n",
" features = pd.merge(features,monbalstats,on='account_id',how='left')\n",
" return features\n",
"\n",
"features = addcols_monbalstats(features,trans,1)\n",
"features = addcols_monbalstats(features,trans,2)\n",
"features = addcols_monbalstats(features,trans,3)\n",
"features = addcols_monbalstats(features,trans,4)\n",
"features = addcols_monbalstats(features,trans,5)\n",
"features = addcols_monbalstats(features,trans,6)\n",
"\n",
"# Quick verification:\n",
"#features[['min1','min2','min3','min4','min5','min6','N1','N2','N3','N4','N5','N6']].head()\n",
"# Good, we see that for a given record (client), N increases and mins decrease with number of months, as they should."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next we must identify our response (target) variable, the loan status. Let's check the counts of the four possible statuses. We'll lump together statuses A and C as \"good\" loan statuses, and B and D as \"bad\" ones. In future we might try redoing the classification using only B (default) or D (delinquency) alone as \"bad\" and see how the model compares. In any case noticing the counts for each of these statuses, we see that we'll need to consider the issue of balancing the data for the classification."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"C 493\n",
"A 258\n",
"D 45\n",
"B 31\n",
"\n",
"Loan status codes:\n",
"A = contract finished, no problems\n",
"B = contract finished, loan not payed\n",
"C = running contract, OK so far\n",
"D = running contract, client in debt\n"
]
}
],
"source": [
"print(features['status'].value_counts().to_string())\n",
"print()\n",
"print(loan.codes)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We now set things up to handle the categorical variables. First let's convert these binary categorical variables to 0s and 1s:"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"# Convert response var `status` = {A,B,C,D} to `response` = {0,1} (AC good, BD bad):\n",
"features.loc[features['status'].isin(['A','C']),'status'] = 1 # good\n",
"features.loc[features['status'].isin(['B','D']),'status'] = 0 # bad\n",
"features.rename(columns = {'status':'response'}, inplace=True) # so binary response now\n",
"\n",
"# Convert gender={M,F} to gender={0,1}:\n",
"features.loc[features['gender'].isin(['F']),'gender'] = 1\n",
"features.loc[features['gender'].isin(['M']),'gender'] = 0\n",
"\n",
"# There are credit card features, but not all clients have cards so these features can be Nan,\n",
"# which isn't acceptable in the modeling. Let's create a `has_card`={0,1} feature, drop the\n",
"# date the card was opened, and then below we'll still use the type_card feature in a way\n",
"# that avoids NaNs.\n",
"features['has_card'] = 0\n",
"features.loc[pd.notnull(features['date_card']),'has_card'] = 1\n",
"features.drop('date_card', axis=1, inplace=True) # since hard to avoid the NaN in the dates"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The rest of the categorical variables have more than two values (ie cardinality > 2), which brings up an interesting issue. Pandas actually has a new `categorical` data type but as of this writing `scikit-learn` doesn't use it; it's a long known/debated issue about the effort required to rewrite the relevant code. But this raises a problem that is especially relevant to tree-based models, such as that we're using below.\n",
"\n",
"A standard solution to encoding multi-valued categorical variables into numerical variables for modeling is \"one-hot encoding\", ie the single categorical variable gets replaced with a series of binary variables, one for each category. However, the sparsity of the resulting dataset increases with the number of categories, and in the present dataset we have categorical variables `district_id_bank` and `district_id_client` (neighborhoods of the bank branches and clients' homes) which have 77 categories! In tree-based models that severe sparsity maps into greatly reduced sensitivity or importance of those variables in the splitting, and those variables will always fall to the bottom of the importance list. The real solution is to use a modeling framework that can handle categorical variables directly (eg H2O, or R). But for here I'm going to just drop the `district_` features (perhaps I could justify it as a choice to actively prevent unfair/illegal discrimination between clients). Technically this issue does affect the other lower-cardinality categorical variables we'll encode too, but not nearly as dramatically; we'll bear this in mind when interpreting the results.\n",
"\n",
"Some relevant links on the issue:\n",
"* https://roamanalytics.com/2016/10/28/are-categorical-variables-getting-lost-in-your-random-forests\n",
"* https://github.com/scikit-learn/scikit-learn/pull/3346\n",
"* https://medium.com/tech-vision/random-forest-classification-with-h2o-python-for-beginners-b31f6e4ccf3c\n",
"* https://github.com/h2oai/h2o-tutorials/blob/master/tutorials/gbm-randomforest/GBM_RandomForest_Example.py"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"# Use pd.get_dummies() to convert categorical vars to binary columns; then drop the original categorical vars:\n",
"catvars = ['stmt_frq','type_disp','type_card'] # 'district_id_bank','district_id_client']\n",
"# Perhaps add 'duration' to catvars too? It's directional but only discrete values {12,24,36,48,60}.\n",
"features = pd.concat([features, pd.get_dummies(features[catvars], prefix=catvars)], axis=1)\n",
"features.drop(catvars,1,inplace=True)\n",
"\n",
"# dropping district features even though they're not included in catvars:\n",
"features.drop(['district_id_bank','district_id_client'],1,inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next we must drop the variables that may meaninglessly confound with the response (eg id/key vars):"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"# Drop confounding variables\n",
"confvars = ['account_id','disp_id','client_id','card_id','loan_id']\n",
"features.drop(confvars,1,inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Lastly, we must convert the date fields to their equivalent UNIX time - an ordered, integer representation which the model can use:"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"features.date_acct = features.date_acct.view('int64')\n",
"features.date_loan = features.date_loan.view('int64')\n",
"features.date_birth = features.date_birth.view('int64')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Great. Now we have 827 records with a `response` column and 34 feature columns in our dataframe, and we're ready to train and analyze a classifier using them."
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"(827, 35)"
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"features.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Classification <a class=\"anchor\" id=\"classification\"></a><small>(return to [top](#top))</small>"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now we look at the balance of the `response` variable in the data. If severely unbalanced (as hopefully it is for real bank data, ie many more good loans than bad), we must do something to address it which may affect the amount of training data. But the amount of training data limits the complexity of the resulting model and what rules can be gleaned from the data governing client behavior re loans. So let's quick check that situation about the rows of training data:"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"827 clients with a loan - one client per feature record\n",
"\n",
"Checking how balanced the loan statuses are for using them as training data:\n",
"1 751\n",
"0 76\n"
]
}
],
"source": [
"# number of account/client pairs (ie feature rows) with a loan_id:\n",
"print(len(features),'clients with a loan - one client per feature record')\n",
"print()\n",
"print('Checking how balanced the loan statuses are for using them as training data:')\n",
"print(features['response'].value_counts().to_string())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So the training data is in fact fairly unbalanced (with a 1:10 ratio), strongly weighted to good loans (again a good thing in reality, although surprisingly high percentage actually). We'll compare effects of balancing the training data via class-weight balancing (built in to SciKit-Learn's RF classifier), the SMOTE synthetic data producing balancing algorithm via the `imbalanced-learn` package from `contrib.scikit-learn`, undersampling the majority class (with `imbalanced-learn`'s RandomUnderSampler class), and doing no balancing at all.\n",
"After that a major question will be to discover which features among the many we have are the most important ones; fortunately Random Forests are good at solving that problem on a relatively small number of data."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Calling my own wrapper encapsulating SciKit-Learn's RandomForestClassifier and scoring methods, and Imbalanced-Learn's resamplers, let's compare effects on multiple shuffled-stratifications. In all case I use a 33%/67% split for test/train data, stratified so that training and test data retain the same class ratio (approx 1:10) between bad and good loans as is in the full dataset:"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Precision Recall F1 ROCarea Accuracy ____ConfusionMatrix____ #trees rebal 0s/1s 0s/1s\n",
"-----------------------------------------------------------------------------------------------------------------\n",
"Pre=0.93 Rec=1.00 F1=0.96 AUC=0.88 Acc=0.93 TP=248 TN=7 FP=18 FN=0 n=100 cwbal train:51/503 test:25/248\n",
"Pre=0.94 Rec=1.00 F1=0.96 AUC=0.90 Acc=0.93 TP=247 TN=8 FP=17 FN=1 n=100 cwbal train:51/503 test:25/248\n",
"Pre=0.94 Rec=0.99 F1=0.96 AUC=0.90 Acc=0.93 TP=246 TN=9 FP=16 FN=2 n=100 cwbal train:51/503 test:25/248\n",
"Pre=0.94 Rec=1.00 F1=0.96 AUC=0.87 Acc=0.93 TP=247 TN=8 FP=17 FN=1 n=100 cwbal train:51/503 test:25/248\n",
" \n",
"Pre=0.95 Rec=0.94 F1=0.94 AUC=0.83 Acc=0.89 TP=232 TN=12 FP=13 FN=16 n=100 smote train:503/503 test:25/248\n",
"Pre=0.96 Rec=0.95 F1=0.96 AUC=0.89 Acc=0.92 TP=235 TN=16 FP=9 FN=13 n=100 smote train:503/503 test:25/248\n",
"Pre=0.96 Rec=0.97 F1=0.96 AUC=0.94 Acc=0.93 TP=240 TN=15 FP=10 FN=8 n=100 smote train:503/503 test:25/248\n",
"Pre=0.96 Rec=0.94 F1=0.95 AUC=0.92 Acc=0.91 TP=234 TN=14 FP=11 FN=14 n=100 smote train:503/503 test:25/248\n",
" \n",
"Pre=0.95 Rec=0.83 F1=0.89 AUC=0.84 Acc=0.81 TP=206 TN=15 FP=10 FN=42 n=100 under train:51/51 test:25/248\n",
"Pre=0.98 Rec=0.74 F1=0.84 AUC=0.88 Acc=0.75 TP=184 TN=21 FP=4 FN=64 n=100 under train:51/51 test:25/248\n",
"Pre=0.97 Rec=0.73 F1=0.83 AUC=0.86 Acc=0.74 TP=181 TN=20 FP=5 FN=67 n=100 under train:51/51 test:25/248\n",
"Pre=0.97 Rec=0.75 F1=0.85 AUC=0.82 Acc=0.75 TP=186 TN=19 FP=6 FN=62 n=100 under train:51/51 test:25/248\n",
" \n",
"Pre=0.94 Rec=1.00 F1=0.97 AUC=0.85 Acc=0.94 TP=248 TN=9 FP=16 FN=0 n=100 None train:51/503 test:25/248\n",
"Pre=0.94 Rec=0.99 F1=0.96 AUC=0.85 Acc=0.93 TP=245 TN=8 FP=17 FN=3 n=100 None train:51/503 test:25/248\n",
"Pre=0.93 Rec=0.98 F1=0.96 AUC=0.85 Acc=0.92 TP=244 TN=8 FP=17 FN=4 n=100 None train:51/503 test:25/248\n",
"Pre=0.95 Rec=0.99 F1=0.97 AUC=0.87 Acc=0.94 TP=246 TN=11 FP=14 FN=2 n=100 None train:51/503 test:25/248\n"
]
}
],
"source": [
"import implement_rf\n",
"print('Precision Recall F1 ROCarea Accuracy ____ConfusionMatrix____ #trees rebal 0s/1s 0s/1s')\n",
"print('-----------------------------------------------------------------------------------------------------------------')\n",
"for i in range(4):\n",
" scores,rc,fimp,roc,clf = implement_rf.implement_sklrf(features, rebal='cwbal', textout_fullsummary=True)\n",
"print(' ')\n",
"for i in range(4):\n",
" scores,rc,fimp,roc,clf = implement_rf.implement_sklrf(features, rebal='smote', textout_fullsummary=True)\n",
"print(' ')\n",
"for i in range(4):\n",
" scores,rc,fimp,roc,clf = implement_rf.implement_sklrf(features, rebal='under', textout_fullsummary=True)\n",
"print(' ')\n",
"for i in range(4):\n",
" scores,rc,fimp,roc,clf = implement_rf.implement_sklrf(features, rebal=None, textout_fullsummary=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Frankly they all do pretty well; this is the first hint that certain features picked out from the transactions data are very sensitive to the loan behavior. But it is interesting to note that the class_weight balancing (`cwbal`) had little effect relative to no balancing at all (`None`). Undersampling the majority class, the most intuitive approach, appears to have degraded performance by most metrics compared to the methods, presumably due to the reduced training data.\n",
"SMOTE appears to've allowed greater true negative resolution (due to more negatives in the training data), at the cost of some more false negatives and very slight reduction in true positive rate. Let's go with the SMOTE balancing for the rest:"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Pre=0.96 Rec=0.92 F1=0.94 AUC=0.84 Acc=0.90 TP=229 TN=16 FP=9 FN=19 n=100 smote train:503/503 test:25/248\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA0QAAAFmCAYAAABN4lo6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x116eef668>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"scores,rc,fimp,roc,clf = implement_rf.implement_sklrf(features, rebal='smote', do_plots=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So there's what we were after in this work - what's the accuracy of this statistical model and what are the important features? With imbalanced data, the accuracy is not typically the metric to focus on, but rebalancing remedies that. The other metrics all show good performance (and repeatably so with every test/train resplit. What really pops out in rerunning the plots above on different splits is how consistently the minimum balance features remain at the front as the most important model features, which makes sense intuitively.\n",
"\n",
"I do see that many of the categorical variables (max cardinality = 3) that I one-hot-encoded into binary variables have ended up at the bottom of the feature importances, and that brings back to mind the discussion earlier about this. It's hard to know how much of an issue this is - max cardinality is 3 but these binary variables form 25% of the features. On the other hand it seems reasonable for things like statement frequency to not be very predictive of loan behavior. The positioning of those variables is suspicious though, and the likely next issue to resolve in this work.\n",
"\n",
"So we see in the feature importances that minimum balances and some others come into play as important... but how exactly? We can find that out in looking at a bunch of the individual decision trees from the forest. (The random forest's result is an average across these trees, so we must look thru many of them.)"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"implement_rf.create_tree_plots(clf, fimp.fnames, ['bad_loan','good_loan'], N=4, max_depth=1)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"<table align=\"left\" width=\"90%\"><tr>\n",
"<td width=\"20%\"><img src=\"tree_0.png\"></td>\n",
"<td width=\"20%\"><img src=\"tree_1.png\"></td>\n",
"<td width=\"20%\"><img src=\"tree_2.png\"></td>\n",
"<td width=\"20%\"><img src=\"tree_3.png\"></td>\n",
"</tr></table>"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The above are only four of the 100 trees, and only the top of the trees for space reasons. But we can see that it's not simply a matter of trees splitting at zero-balances - rather in scanning over many of these trees they split on \"low\" balances in the months preceding the loan application.\n",
"\n",
"We can gain some further intuition on the results by exploring histograms of the different features, conditioned on the response variable. I.e. let's make histograms of the features when response was bad and compare them to histograms of the features when the response was good."
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"# Convert the date vars back to \"datetime\" format first so it makes sense in the histograms\n",
"features['date_acct'] = pd.to_datetime(features['date_acct'])\n",
"features['date_loan'] = pd.to_datetime(features['date_loan'])\n",
"features['date_birth'] = pd.to_datetime(features['date_birth'])"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA3oAAADSCAYAAAAc9LJmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAaY9ADAABojEEPAACgMXfU1EyqPN2K5QAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x1170bc6d8>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Plot histograms for the 8 strongest features for good and bad loan cases\n",
"tmp = fimp.sort_values(by='fimport',ascending=False)['fnames'][:4]\n",
"tmp.loc[-1]='response'\n",
"fig, axes = plt.subplots(1, 4, figsize=(15,3))\n",
"for i,ax in enumerate(axes.flatten()):\n",
" features[tmp].groupby('response')[tmp.iloc[i]].hist(alpha=0.4,ax=ax)\n",
" ax.set_title(tmp.iloc[i])\n",
" if i==0:\n",
" ax.legend(['bad','good'])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We immediately notice the unbalanced data in these histograms - there's way more \"good\" loans data (green) than \"bad\" loans data (lavender). But aside from that, what is striking in these histograms for the strongest features, is the peak at near-zero balance for the bad loans, which is not seen in the good loans histograms. In fact the good loans have peaks at much higher balance.\n",
"\n",
"Now for inter-feature relationships (like correlation), just looking at scatter plots..."
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"# (alas scatter plot won't handle datetime types, so we convert back to int again)\n",
"features.date_acct = features.date_acct.view('int64')\n",
"features.date_loan = features.date_loan.view('int64')\n",
"features.date_birth = features.date_birth.view('int64')"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA5MAAADeCAYAAACgwVSmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmNfuAAAAAElFTkSuQmCC\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x1172f5a58>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"fig, ax = plt.subplots(1,4,figsize=(15,3))\n",
"features.plot(kind='scatter',x='min1',y='min2',alpha=0.2,grid=True,ax=ax[0],xlim=(-1e4,8e4),ylim=(-1e4,8e4))\n",
"features.plot(kind='scatter',x='min1',y='min6',alpha=0.2,grid=True,ax=ax[1],xlim=(-1e4,8e4),ylim=(-1e4,8e4))\n",
"features.plot(kind='scatter',x='date_loan',y='date_acct',alpha=0.2,grid=True,ax=ax[2],xlim=(7.2e17,9.3e17),ylim=(7.2e17,8.9e17))\n",
"features.plot(kind='scatter',x='amount',y='payments',alpha=0.2,grid=True,ax=ax[3],xlim=(-1e4,6.5e5),ylim=(-1e2,1.1e4))\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Some interesting things I noticed in the scatter plots are (and just plotting these instead of all of them):\n",
"* You can see the zero balances pop out in `min5` and `min6` variables, as horizontal or vertical lines within the spread, but not in the `min1` to `min4` features.\n",
"* Also in the `min` variables, the diagonal-cutoff form of them validates that `min1 < min2 < min3 < min4 < min5 < min6`.\n",
"* The minimum balances are correlated, more so the closer the month spans are to each other (so e.g. note above that min1 vs min2 more closerly correlated than min1 vs min6).\n",
"* The band in the histogram for `date_acct` (date account opened) and `date_loan` (date of loan) shows a ~15 month limit on the span between the two dates, presumably caused by an intentional constraint in querying the original data -- ie \"extract clients for whom the loan date was no later than ~15 months after the account-opening date -- and perhaps highlights that the original dataset was created with loan themes in mind.\n",
"* The monthly `payments` and loan `amount` variables obviously are extremely closely correlated; those five lines in the histogram are the five loan rates corresponding to the five loan durations in the dataset. Oops - redundant information there that suggests a feature reduction in next version."
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"### Next steps... <a class=\"anchor\" id=\"next\"></a><small>(return to [top](#top))</small>"
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"* Explore (for bugs/misunderstandings) the pattern I notice that many of the categorical variables went to the bottom of the feature importances - that seems suspicious to me, given the discussion about one-hotted categorical variables in random forests. If this is found to be something to fix, will that have a significant effect on the conclusions about what the strongest features are?\n",
"* Adding in some other features, like district_id_bank and district_id_client, or even all the linked regional/demographic information within the district_id table for those _bank and _client ids.\n",
"* Try redoing the classification using only B (default) or D (delinquency) alone as \"bad\" and see how the model compares.\n",
"* A different quantity we could try to predict is whether a client is likely to overdraw based on their earlier transaction history plus all the other features..."
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"#### Website links...\n",
"* [over-sampling for unbalanced data](http://contrib.scikit-learn.org/imbalanced-learn/stable/auto_examples/applications/plot_over_sampling_benchmark_lfw.html)\n",
"* [imbalanced-learn user guide on contrib.scikit-learn.org](http://contrib.scikit-learn.org/imbalanced-learn/stable/user_guide.html)\n",
"* [RF importances](http://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html#sphx-glr-auto-examples-ensemble-plot-forest-importances-py)\n",
"* [RF on small datasets](https://stats.stackexchange.com/questions/192310/is-random-forest-suitable-for-very-small-data-sets) \n",
"* [SKL Ensemble methods page, at RF section](http://scikit-learn.org/stable/modules/ensemble.html#forest)\n",
"* [SKL Random Forest Classifier API page](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier)\n",
"* [for ideas about what would benefit customers to experiment with as nudges (but note written by a bank)](https://bettermoneyhabits.bankofamerica.com/en/personal-banking/checking-account-information)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python3",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.5.2"
}
},
"nbformat": 4,
"nbformat_minor": 0
}