Amount Withdrawn Model Part 5

TASK 3

target ='ATM RATING'

from sklearn.ensemble import RandomForestClassifier,GradientBoostingClassifier
from sklearn.tree import DecisionTreeClassifier
from xgboost import XGBClassifier
from lightgbm import LGBMClassifier
from sklearn.linear_model import LogisticRegression

from sklearn.metrics import roc_auc_score,accuracy_score


from sklearn.preprocessing import  MinMaxScaler
from sklearn.cluster import KMeans

from sklearn.decomposition import PCA

evaluate_accuracy= make_scorer(accuracy_score)

log_reg = LogisticRegression()
decision_tree_clf =DecisionTreeClassifier()
rf_clf = RandomForestClassifier()
gbm_clf = GradientBoostingClassifier()
xgb_clf = XGBClassifier()
lgb_clf = LGBMClassifier()

# dataset[FSET],dataset[target]

dataset_cat = pd.get_dummies(dataset[exploration_dict['possible_categorical_features']])
dataset_continuos = dataset[continuos_columns]
dataset_fin = dataset_continuos.join(dataset_cat)
FSET=[ col for col in dataset_fin if col not in [target]]


X = dataset_fin[FSET].copy()
y= dataset[target]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)

X_train.join(y_train).corr()[target].sort_values(ascending =False)

ATM RATING                                          1.000000
Avg Withdrawls Per Hour                             0.672877
Avg No of Withdrawls Per Week                       0.387591
Holiday Sequence_Long Weekend                       0.339350
ATM Prox_Attached to Branch                         0.273523
ATM Prox_Other Bank ATM                             0.193117
Day Type_Working                                    0.180383
No of Other ATMs in 1 KM radius                     0.166138
ATM Zone_FV                                         0.158193
ATM TYPE_Town                                       0.148302
ATM Zone_RL                                         0.122217
ATM Attached to_Petrol Bunk                         0.105171
Estimated Number of Houses in 1 KM Radius           0.066883
ATM Placement_Facing Road                           0.063644
ATM TYPE_Semi Urban                                 0.061452
ATM Location TYPE_Only WIthdraw                     0.047002
ATM Location TYPE_Passbook Printing and Withdraw    0.044917
ATM looks_Normal                                    0.029628
ATM TYPE_Bigger Towns                               0.009193
ATM Since                                           0.004059
ATM Attached to_Shed                                0.000008
ATM Location TYPE_Checkdrop and Withdraw           -0.002357
ATM Attached to_Shop                               -0.010299
Number of Shops Around ATM                         -0.011378
ATM looks_New                                      -0.013827
ATM Attached to_Flat                               -0.017295
Day Type_National Holiday                          -0.020809
Average Wait Time                                  -0.021180
ATM Attached to_House                              -0.024414
Day Type_Strike                                    -0.028450
ATM looks_Old and Dull                             -0.040530
Holiday Sequence_HH                                -0.042008
Holiday Sequence_HW                                -0.047163
ATM Prox_2 ATMS in same Building                   -0.057838
ATM Prox_Non Closed ATM                            -0.058048
Holiday Sequence_WH                                -0.059719
ATM Placement_Little Inside                        -0.063644
ATM Zone_RH                                        -0.080320
ATM Zone_C                                         -0.087497
ATM Attached to_Building                           -0.090750
ATM Location TYPE_Deposit and Withdraw             -0.108344
ATM TYPE_Urban                                     -0.167517
Day Type_Festival                                  -0.176559
ATM Zone_RM                                        -0.187803
Holiday Sequence_WW                                -0.201102
ATM Prox_2+ ATMS Very Close                        -0.213559
ATM Prox_2 ATM Machines (Same Bank)                -0.269406
Name: ATM RATING, dtype: float64
model_name = 'log_reg,decision_tree_clf,rf_clf,gbm_clf,xgb_clf,lgb_clf'


for m in model_name.split(','):
    print(m)
    model = eval(m)

    model.fit(X_train[['Avg Withdrawls Per Hour']].values,y_train.values)


    cv_results = cross_val_score(model,X_train.values,y_train.values,cv=10,scoring= evaluate_accuracy)


    preds_y = model.predict(X_test[['Avg Withdrawls Per Hour']].values)


    print(f'Test Results of model {m}')
    print(f'Insample Accuracy:{cv_results.mean()}, Test Accuracy :{accuracy_score(preds_y,y_test)}')

log_reg
Test Results of model log_reg
Insample Accuracy:0.7456810232460609, Test Accuracy :0.7209060984903763
decision_tree_clf
Test Results of model decision_tree_clf
Insample Accuracy:1.0, Test Accuracy :0.7495257090042321
rf_clf
Test Results of model rf_clf
Insample Accuracy:1.0, Test Accuracy :0.7495257090042321
gbm_clf
Test Results of model gbm_clf
Insample Accuracy:0.8626136542309855, Test Accuracy :0.7495257090042321
xgb_clf
Test Results of model xgb_clf
Insample Accuracy:0.825667445991068, Test Accuracy :0.7495257090042321
lgb_clf
Test Results of model lgb_clf
Insample Accuracy:0.9880280092542281, Test Accuracy :0.7495257090042321

model_name = 'log_reg,decision_tree_clf,rf_clf,gbm_clf,xgb_clf,lgb_clf'


for m in model_name.split(','):
    print(m)
    model = eval(m)

    model.fit(X_train[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values,y_train.values)


    cv_results = cross_val_score(model,X_train.values,y_train.values,cv=10,scoring= evaluate_accuracy)


    preds_y = model.predict(X_test[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values)


    print(f'Test Results of model {m}')
    print(f'Insample Accuracy:{cv_results.mean()}, Test Accuracy :{accuracy_score(preds_y,y_test)}')

log_reg
Test Results of model log_reg
Insample Accuracy:0.7456810232460609, Test Accuracy :0.7223005950933178
decision_tree_clf
Test Results of model decision_tree_clf
Insample Accuracy:1.0, Test Accuracy :0.9420959608243745
rf_clf
Test Results of model rf_clf
Insample Accuracy:1.0, Test Accuracy :0.9424851226670559
gbm_clf
Test Results of model gbm_clf
Insample Accuracy:0.8626136542309855, Test Accuracy :0.8095863533913833
xgb_clf
Test Results of model xgb_clf
Insample Accuracy:0.825667445991068, Test Accuracy :0.7849880819185678
lgb_clf
Test Results of model lgb_clf
Insample Accuracy:0.9880280092542281, Test Accuracy :0.8499132493392356

result_df = pd.DataFrame(preds_y,columns =['preds'])
result_df['actuals'] = y_test.values

feat_importance = pd.DataFrame(rf_clf.feature_importances_,columns=['feature_importance'])
feat_importance['features'] = X_train.columns.values

feat_importance.sort_values('feature_importance',ascending = False)

feature_importance features
1 0.229098 Avg Withdrawls Per Hour
3 0.151150 Avg No of Withdrawls Per Week
2 0.132929 Estimated Number of Houses in 1 KM Radius
6 0.098502 No of Other ATMs in 1 KM radius
5 0.055319 Number of Shops Around ATM
4 0.049528 ATM Since
0 0.039271 Average Wait Time
43 0.020507 ATM Prox_Attached to Branch
31 0.019383 Holiday Sequence_Long Weekend
40 0.017415 ATM Prox_2 ATM Machines (Same Bank)
34 0.015416 ATM Attached to_Building
25 0.014983 ATM Zone_FV
15 0.013845 ATM TYPE_Urban
14 0.012845 ATM TYPE_Town
37 0.011269 ATM Attached to_Petrol Bunk
28 0.008970 ATM Zone_RM
33 0.008907 Holiday Sequence_WW
27 0.008863 ATM Zone_RL
18 0.008844 ATM Location TYPE_Only WIthdraw
45 0.007995 ATM Prox_Other Bank ATM
20 0.007241 Day Type_Festival
17 0.006592 ATM Location TYPE_Deposit and Withdraw
29 0.006165 Holiday Sequence_HH
42 0.005072 ATM Prox_2+ ATMS Very Close
23 0.004871 Day Type_Working
10 0.004856 ATM looks_Normal
19 0.004796 ATM Location TYPE_Passbook Printing and Withdraw
9 0.004381 ATM looks_New
35 0.004278 ATM Attached to_Flat
13 0.004117 ATM TYPE_Semi Urban
44 0.004089 ATM Prox_Non Closed ATM
16 0.003192 ATM Location TYPE_Checkdrop and Withdraw
26 0.002621 ATM Zone_RH
30 0.002191 Holiday Sequence_HW
41 0.001815 ATM Prox_2 ATMS in same Building
11 0.001667 ATM looks_Old and Dull
39 0.001509 ATM Attached to_Shop
12 0.001336 ATM TYPE_Bigger Towns
24 0.001053 ATM Zone_C
36 0.000903 ATM Attached to_House
32 0.000775 Holiday Sequence_WH
7 0.000681 ATM Placement_Facing Road
8 0.000349 ATM Placement_Little Inside
22 0.000238 Day Type_Strike
38 0.000161 ATM Attached to_Shed
21 0.000010 Day Type_National Holiday 
for max_depth in np.arange(1,50):
    model = RandomForestClassifier(max_depth=max_depth)
    model.fit(X_train[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values,y_train.values)
    cv_results = cross_val_score(model,X_train[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values,y_train.values,cv=10,scoring= evaluate_accuracy)
    preds_y = model.predict(X_test[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values)
    print(f'Maxdepth: {max_depth}, Insample Accuracy:{cv_results.mean()}, Test Accuracy :{accuracy_score(preds_y,y_test)}')

Maxdepth: 1, Insample Accuracy:0.7224159535638427, Test Accuracy :0.7209060984903763
Maxdepth: 2, Insample Accuracy:0.7446346667171084, Test Accuracy :0.7403155453941075
Maxdepth: 3, Insample Accuracy:0.7458009801722265, Test Accuracy :0.7497040748487944
Maxdepth: 4, Insample Accuracy:0.7523499135498797, Test Accuracy :0.7492014074686644
Maxdepth: 5, Insample Accuracy:0.7678121085241824, Test Accuracy :0.7690648765221904
Maxdepth: 6, Insample Accuracy:0.779017245108138, Test Accuracy :0.7800262684243809
Maxdepth: 7, Insample Accuracy:0.7913646951946998, Test Accuracy :0.7920092101636101
Maxdepth: 8, Insample Accuracy:0.8083041126314837, Test Accuracy :0.8051920675844401
Maxdepth: 9, Insample Accuracy:0.8280713703549882, Test Accuracy :0.8292876716771254
Maxdepth: 10, Insample Accuracy:0.845786067440496, Test Accuracy :0.8426164647889608
Maxdepth: 11, Insample Accuracy:0.8614795990190192, Test Accuracy :0.8577775615767541
Maxdepth: 12, Insample Accuracy:0.8777962710041655, Test Accuracy :0.8750141881921811
Maxdepth: 13, Insample Accuracy:0.8902633724832689, Test Accuracy :0.8880511099220055
Maxdepth: 14, Insample Accuracy:0.9060534383659145, Test Accuracy :0.8989800716706393
Maxdepth: 15, Insample Accuracy:0.9143031655480852, Test Accuracy :0.9036986590131504
Maxdepth: 16, Insample Accuracy:0.9201736711788391, Test Accuracy :0.9291887597087772
Maxdepth: 17, Insample Accuracy:0.9282078308355667, Test Accuracy :0.9316858815326491
Maxdepth: 18, Insample Accuracy:0.9327604878738853, Test Accuracy :0.9240323652932496
Maxdepth: 19, Insample Accuracy:0.9354438654358512, Test Accuracy :0.9301940944690373
Maxdepth: 20, Insample Accuracy:0.9402759138242723, Test Accuracy :0.940863614989217
Maxdepth: 21, Insample Accuracy:0.9409306458414276, Test Accuracy :0.9413176371390118
Maxdepth: 22, Insample Accuracy:0.9424083148070034, Test Accuracy :0.9428580694329588
Maxdepth: 23, Insample Accuracy:0.9424961767864005, Test Accuracy :0.9429391448168507
Maxdepth: 24, Insample Accuracy:0.9427678098951573, Test Accuracy :0.9428418543561804
Maxdepth: 25, Insample Accuracy:0.942911491844718, Test Accuracy :0.9423716171296072
Maxdepth: 26, Insample Accuracy:0.9423763997344338, Test Accuracy :0.9429229297400723
Maxdepth: 27, Insample Accuracy:0.9429674349173849, Test Accuracy :0.9425661980509478
Maxdepth: 28, Insample Accuracy:0.942368378706601, Test Accuracy :0.9425661980509478
Maxdepth: 29, Insample Accuracy:0.9425361256466539, Test Accuracy :0.9417554442120284
Maxdepth: 30, Insample Accuracy:0.9428875180661175, Test Accuracy :0.9424202623599423
Maxdepth: 31, Insample Accuracy:0.9423763506169995, Test Accuracy :0.9425175528206126
Maxdepth: 32, Insample Accuracy:0.9425600075741322, Test Accuracy :0.9423554020528288
Maxdepth: 33, Insample Accuracy:0.9420169818259735, Test Accuracy :0.9425986282045046
Maxdepth: 34, Insample Accuracy:0.9422326036169251, Test Accuracy :0.9426634885116181
Maxdepth: 35, Insample Accuracy:0.9420648732534611, Test Accuracy :0.94224189651538
Maxdepth: 36, Insample Accuracy:0.9424722017314779, Test Accuracy :0.9423716171296072
Maxdepth: 37, Insample Accuracy:0.942288552427995, Test Accuracy :0.9423554020528288
Maxdepth: 38, Insample Accuracy:0.9421926770771651, Test Accuracy :0.942533767897391
Maxdepth: 39, Insample Accuracy:0.9420249856346243, Test Accuracy :0.9424364774367207
Maxdepth: 40, Insample Accuracy:0.94231248729637, Test Accuracy :0.9426472734348397
Maxdepth: 41, Insample Accuracy:0.9425600330886379, Test Accuracy :0.9424689075902775
Maxdepth: 42, Insample Accuracy:0.9427437487278212, Test Accuracy :0.94224189651538
Maxdepth: 43, Insample Accuracy:0.9422006694028978, Test Accuracy :0.9423878322063854
Maxdepth: 44, Insample Accuracy:0.9428954536174041, Test Accuracy :0.9425175528206126
Maxdepth: 45, Insample Accuracy:0.9428315966744071, Test Accuracy :0.942533767897391
Maxdepth: 46, Insample Accuracy:0.9420649223708961, Test Accuracy :0.9424689075902775
Maxdepth: 47, Insample Accuracy:0.9423603838361723, Test Accuracy :0.9424040472831638
Maxdepth: 48, Insample Accuracy:0.9421847160116771, Test Accuracy :0.9423391869760503
Maxdepth: 49, Insample Accuracy:0.9427996503409819, Test Accuracy :0.9423067568224935

model = RandomForestClassifier()
model.fit(X_train[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values,y_train.values)
cv_results = cross_val_score(model,X_train[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values,y_train.values,cv=10,scoring= evaluate_accuracy)
preds_y = model.predict(X_test[['Avg Withdrawls Per Hour','Avg No of Withdrawls Per Week']].values)
print(f'Insample Accuracy:{cv_results.mean()}, Test Accuracy :{accuracy_score(preds_y,y_test)}')

Insample Accuracy:0.9423444419242031, Test Accuracy :0.9423554020528288

result_df = pd.DataFrame(preds_y,columns =['preds'])
result_df['actuals'] = y_test.values
display(pd.crosstab(result_df['actuals'],result_df['preds']))
preds 2 3 4 5
actuals



2 1297 328 42 0
3 43 29766 1216 0
4 100 1402 23174 75
5 0 130 219 3879

Comments

Post a Comment

Popular posts from this blog

Amount Withdrawn Model Part 2

Image
Univariate visualization features = exploration_dict [ 'possible_continuos_features' ] dataset [ features ] . hist ( figsize = ( 20 , 20 )); dataset [ features ] . plot ( kind = 'density' , subplots = True , layout = ( 4 , 3 ), sharex = False , figsize = ( 20 , 20 )); # Sometimes you can analyze an ordinal variable just as numerical one fig , axes = plt . subplots ( nrows = 3 , ncols = 4 , figsize = ( 30 , 30 )) for idx , feat in enumerate ( features ): ax = axes [ int ( idx / 4 ), idx % 4 ] sns . boxplot ( x = feat , data = dataset , ax = ax ) ax . set_xlabel ( feat ) ax . set_ylabel ( '' ) fig . tight_layout (); from pandas.plotting import scatter_matrix % config InlineBackend.figure_format = 'png' style . use ( 'tableau-colorbli...
Read more

Hybrid Recommendation Engine

Hybrid Recommender System Intro In this project we are building a Hybrid recommender system which is a combination of 1. Content Based Filtering (Tag Based)  2. Collaborative Filtering Recommender System. The goal of this exercise is to recommend an item for a particular user based on the above recommender systems. We are going particularly implement the following paper in the link Work Flow of the Recommender System Like we mentioned the proposed system is comprised of 2 parts:  1. Data set Preparation 2. Collaborative Filtering Stage a. Creating User Profile based on User Related Tags b. Perform Clustering Based on Similarity Behind the Users  3. Content Based Filtering Stage a. Users from each cluster has similar preferences, so we perform Content Based Recommender within each cluster. b. Creating Item Profiles c. Perform Cosine Similarity Clustering between User Profiles and Item Profiles d. Calculate Tag Recency ...
Read more

Amount Withdrawn Model Part 1

Image
% load_ext autoreload % autoreload 2 import sys sys . path . insert( 0 , '../../Models/ALS' ) from stark import * ;hello() / home / ec2 - user / SageMaker / SUNDARESAN M S / Scripts / HandsOnML - CookBook Data Science Environment Set ! from sklearn.model_selection import train_test_split from sklearn.model_selection import cross_val_score,KFold from sklearn.pipeline import make_pipeline from sklearn.metrics import mean_absolute_error,mean_squared_error from sklearn.metrics import make_scorer from sklearn.linear_model import ElasticNet, Lasso from sklearn.linear_model import LinearRegression from sklearn.tree import DecisionTreeRegressor from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor from xgboost import XGBRegressor from lightgbm import LGBMRegressor from matplotlib.pyplot import style import matplotlib.pyplot as plt import seaborn as sns import pandas as pd style . use( '...
Read more