1. Imports

import pandas as pd
import numpy as np
import sklearn
import pickle 
import time 
import datetime

import warnings
warnings.filterwarnings('ignore')

%run ../functions_pyod2.py

with open('../fraudTrain.pkl', 'rb') as file:
    fraudTrain = pickle.load(file)

2.

pyod(*pyod_preprocess4(fraudTrain, 0.03))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.030764	0.963237	0.386093	0.389706	0.387891	0.685307	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
1	GMM	0.029831	0.969790	0.494832	0.512032	0.503285	0.747963	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
2	HBOS	0.007478	0.969750	0.000000	0.000000	0.000000	0.499815	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
3	IForest	3.284466	0.969790	0.494832	0.512032	0.503285	0.747963	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
4	INNE	3.445417	0.969790	0.494832	0.512032	0.503285	0.747963	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
5	KNN	0.179364	0.969910	0.496762	0.512701	0.504605	0.748349	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
6	LODA	0.237153	0.969750	0.000000	0.000000	0.000000	0.499815	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
7	LOF	0.298528	0.933566	0.026411	0.034091	0.029764	0.497686	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
8	MCD	0.027775	0.969790	0.494832	0.512032	0.503285	0.747963	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
9	PCA	0.009329	0.968591	0.474901	0.480615	0.477741	0.732121	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None
10	ROD	6.908336	0.940480	0.000673	0.000668	0.000671	0.485052	False	pyod	0.03	150150	[amt]	0.030037	50050	0.02989	None

pyod(*pyod_preprocess4(fraudTrain, 0.02))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.042985	0.973533	0.346720	0.338784	0.342706	0.662756	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
1	GMM	0.039798	0.979873	0.506081	0.489863	0.497840	0.739962	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
2	HBOS	0.010284	0.979261	0.000000	0.000000	0.000000	0.499810	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
3	IForest	4.639415	0.980100	0.511978	0.489209	0.500334	0.739757	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
4	INNE	5.045339	0.980420	0.520673	0.485939	0.502706	0.738319	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
5	KNN	0.286477	0.977955	0.456671	0.434271	0.445189	0.711765	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
6	LODA	0.374500	0.979261	0.000000	0.000000	0.000000	0.499810	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
7	LOF	0.506328	0.958495	0.002508	0.002616	0.002561	0.490492	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
8	MCD	0.037995	0.979873	0.506081	0.489863	0.497840	0.739962	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
9	PCA	0.013600	0.979660	0.500681	0.480706	0.490490	0.735370	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None
10	ROD	5.371442	0.967939	0.000000	0.000000	0.000000	0.494031	False	pyod	0.02	225225	[amt]	0.019878	75075	0.020366	None

pyod(*pyod_preprocess4(fraudTrain, 0.01))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.083728	0.984695	0.243837	0.239372	0.241584	0.615868	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
1	GMM	0.072678	0.989331	0.475715	0.467626	0.471636	0.731162	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
2	HBOS	0.017917	0.989457	0.000000	0.000000	0.000000	0.499818	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
3	IForest	10.094498	0.989351	0.476667	0.467626	0.472103	0.731172	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
4	INNE	11.064828	0.989404	0.479223	0.467626	0.473353	0.731199	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
5	KNN	0.667739	0.987686	0.394040	0.389143	0.391576	0.691493	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
6	LODA	0.684057	0.989457	0.000000	0.000000	0.000000	0.499818	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
7	LOF	1.079839	0.980773	0.002199	0.001962	0.002074	0.496402	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
8	MCD	0.083876	0.989331	0.475715	0.467626	0.471636	0.731162	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
9	PCA	0.029059	0.989331	0.475715	0.467626	0.471636	0.731162	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None
10	ROD	10.545987	0.975425	0.000462	0.000654	0.000542	0.493053	False	pyod	0.01	450450	[amt]	0.009939	150150	0.010183	None

pyod(*pyod_preprocess4(fraudTrain, 0.00573))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.157627	0.989803	0.118100	0.124157	0.121053	0.559442	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
1	GMM	0.144377	0.992887	0.376615	0.393387	0.384818	0.694842	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
2	HBOS	0.023745	0.993982	0.000000	0.000000	0.000000	0.499818	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
3	IForest	20.173736	0.992887	0.376615	0.393387	0.384818	0.694842	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
4	INNE	21.359880	0.989490	0.053371	0.051282	0.052306	0.523054	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
5	KNN	1.404509	0.992116	0.303235	0.303644	0.303439	0.649838	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
6	LODA	1.521674	0.993982	0.000000	0.000000	0.000000	0.499818	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
7	LOF	2.060348	0.988952	0.001411	0.001350	0.001380	0.497959	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
8	MCD	0.135239	0.992887	0.376615	0.393387	0.384818	0.694842	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
9	PCA	0.047284	0.992887	0.376615	0.393387	0.384818	0.694842	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None
10	ROD	11.951331	0.990101	0.000000	0.000000	0.000000	0.497866	False	pyod	0.00573	786126	[amt]	0.005755	262042	0.005656	None

pyod(*pyod_preprocess4(fraudTrain, 0.006))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.145698	0.989686	0.122661	0.118633	0.120613	0.556772	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
1	GMM	0.135238	0.992619	0.374912	0.356568	0.365510	0.676501	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
2	HBOS	0.024256	0.993758	0.000000	0.000000	0.000000	0.499859	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
3	IForest	16.434630	0.992599	0.371977	0.350536	0.360939	0.673493	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
4	INNE	19.408525	0.992715	0.372985	0.325737	0.347764	0.661226	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
5	KNN	1.251204	0.991708	0.300752	0.294906	0.297800	0.645397	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
6	LODA	1.181331	0.993758	0.000000	0.000000	0.000000	0.499859	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
7	LOF	1.943183	0.988024	0.001325	0.001340	0.001333	0.497641	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
8	MCD	0.132324	0.992619	0.374912	0.356568	0.365510	0.676501	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
9	PCA	0.042269	0.992619	0.374912	0.356568	0.365510	0.676501	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None
10	ROD	13.854177	0.993958	0.000000	0.000000	0.000000	0.499960	False	pyod	0.006	750750	[amt]	0.006013	250250	0.005962	None

pyod(*pyod_preprocess4(fraudTrain, 0.007))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.127007	0.988061	0.159015	0.155599	0.157289	0.574832	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
1	GMM	0.118660	0.991855	0.428475	0.411458	0.419794	0.703750	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
2	HBOS	0.021492	0.992499	0.000000	0.000000	0.000000	0.499829	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
3	IForest	13.654450	0.991828	0.425736	0.404948	0.415082	0.700504	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
4	INNE	16.246222	0.991930	0.428044	0.377604	0.401245	0.686983	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
5	KNN	1.089241	0.990247	0.311141	0.298177	0.304521	0.646708	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
6	LODA	1.013922	0.992499	0.000000	0.000000	0.000000	0.499829	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
7	LOF	1.578532	0.987096	0.001618	0.001302	0.001443	0.497754	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
8	MCD	0.110214	0.991855	0.428475	0.411458	0.419794	0.703750	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
9	PCA	0.036325	0.991855	0.428475	0.411458	0.419794	0.703750	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None
10	ROD	12.376851	0.988573	0.000000	0.000000	0.000000	0.497852	False	pyod	0.007	643500	[amt]	0.006946	214500	0.007161	None

pyod(*pyod_preprocess4(fraudTrain, 0.008))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.107724	0.987197	0.197100	0.201482	0.199267	0.597471	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
1	GMM	0.101690	0.991156	0.442857	0.459569	0.451058	0.727480	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
2	HBOS	0.016904	0.991710	0.000000	0.000000	0.000000	0.499807	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
3	IForest	14.253869	0.991182	0.443861	0.455526	0.449618	0.725488	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
4	INNE	15.132947	0.991193	0.444298	0.454178	0.449184	0.724825	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
5	KNN	0.851337	0.989728	0.357875	0.376685	0.367039	0.685649	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
6	LODA	0.876604	0.991710	0.000000	0.000000	0.000000	0.499807	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
7	LOF	1.358122	0.984964	0.000000	0.000000	0.000000	0.496407	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
8	MCD	0.095074	0.991156	0.442857	0.459569	0.451058	0.727480	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
9	PCA	0.033198	0.991156	0.442857	0.459569	0.451058	0.727480	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None
10	ROD	11.736433	0.987858	0.000000	0.000000	0.000000	0.497865	False	pyod	0.008	563062	[amt]	0.008031	187688	0.007907	None

pyod(*pyod_preprocess4(fraudTrain, 0.009))

	model	time	acc	pre	rec	f1	auc	graph_based	method	throw_rate	train_size	train_cols	train_frate	test_size	test_frate	hyper_params
0	ECOD	0.099683	0.985872	0.220540	0.222149	0.221341	0.607493	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
1	GMM	0.093896	0.990164	0.456106	0.458223	0.457162	0.726619	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
2	HBOS	0.018281	0.990583	0.000000	0.000000	0.000000	0.499809	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
3	IForest	11.115221	0.990350	0.465306	0.453581	0.459369	0.724413	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
4	INNE	12.340797	0.990134	0.454605	0.458223	0.456407	0.726604	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
5	KNN	0.748844	0.988797	0.379426	0.376658	0.378037	0.685519	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
6	LODA	1.080885	0.990583	0.000000	0.000000	0.000000	0.499809	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
7	LOF	1.217433	0.982210	0.002046	0.001989	0.002017	0.496570	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
8	MCD	0.093267	0.990164	0.456106	0.458223	0.457162	0.726619	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
9	PCA	0.037229	0.990164	0.456106	0.458223	0.457162	0.726619	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None
10	ROD	12.339864	0.977996	0.000000	0.000000	0.000000	0.493458	False	pyod	0.009	500499	[amt]	0.008987	166834	0.009039	None