keras-io/imbalanced_classification

Model Description

Keras Implementation of Imbalanced classification: credit card fraud detection

This repo contains the trained model of Imbalanced classification: credit card fraud detection. The full credit goes to: fchollet

Intended uses & limitations

• The trained model is used to detect of a specific transaction is fraudulent or not.

Training dataset

• Credit Card Fraud Detection
• Due to the high imbalance of the target feature (417 frauds or 0.18% of total 284,807 samples), training weight was applied to reduce the False Negatives to the lowest level as possible.

Training procedure

Training hyperparameter

The following hyperparameters were used during training:

• optimizer: 'Adam'
• learning_rate: 0.01
• loss: 'binary_crossentropy'
• epochs: 30
• batch_size: 2048
• beta_1: 0.9
• beta_2: 0.999
• epsilon: 1e-07
• training_precision: float32

Training Metrics

Epochs	Train Loss	Train Fn	Train Fp	Train Tn	Train Tp	Train Precision	Train Recall	Validation Loss	Validation Fn	Validation Fp	Validation Tn	Validation Tp	Validation Precision	Validation Recall
1	0.0	14.0	6202.0	221227.0	403.0	0.061	0.966	0.043	9.0	622.0	56264.0	66.0	0.096	0.88
2	0.0	3.0	3514.0	223915.0	414.0	0.105	0.993	0.025	10.0	528.0	56358.0	65.0	0.11	0.867
3	0.0	2.0	2419.0	225010.0	415.0	0.146	0.995	0.014	11.0	283.0	56603.0	64.0	0.184	0.853
4	0.0	3.0	2482.0	224947.0	414.0	0.143	0.993	0.027	11.0	340.0	56546.0	64.0	0.158	0.853
5	0.0	2.0	2295.0	225134.0	415.0	0.153	0.995	0.034	11.0	245.0	56641.0	64.0	0.207	0.853
6	0.0	3.0	2239.0	225190.0	414.0	0.156	0.993	0.037	10.0	495.0	56391.0	65.0	0.116	0.867
7	0.0	2.0	3095.0	224334.0	415.0	0.118	0.995	0.011	11.0	194.0	56692.0	64.0	0.248	0.853
8	0.0	4.0	1844.0	225585.0	413.0	0.183	0.99	0.035	9.0	429.0	56457.0	66.0	0.133	0.88
9	0.0	1.0	2119.0	225310.0	416.0	0.164	0.998	0.012	11.0	167.0	56719.0	64.0	0.277	0.853
10	0.0	3.0	1539.0	225890.0	414.0	0.212	0.993	0.013	13.0	144.0	56742.0	62.0	0.301	0.827
11	0.0	6.0	3444.0	223985.0	411.0	0.107	0.986	0.039	11.0	394.0	56492.0	64.0	0.14	0.853
12	0.0	4.0	3818.0	223611.0	413.0	0.098	0.99	0.03	9.0	523.0	56363.0	66.0	0.112	0.88
13	0.0	7.0	4482.0	222947.0	410.0	0.084	0.983	0.059	6.0	1364.0	55522.0	69.0	0.048	0.92
14	0.0	2.0	3064.0	224365.0	415.0	0.119	0.995	0.033	9.0	699.0	56187.0	66.0	0.086	0.88
15	0.0	4.0	3563.0	223866.0	413.0	0.104	0.99	0.066	8.0	956.0	55930.0	67.0	0.065	0.893
16	0.0	4.0	2536.0	224893.0	413.0	0.14	0.99	0.016	9.0	339.0	56547.0	66.0	0.163	0.88
17	0.0	6.0	2594.0	224835.0	411.0	0.137	0.986	0.049	8.0	821.0	56065.0	67.0	0.075	0.893
18	0.0	1.0	1911.0	225518.0	416.0	0.179	0.998	0.013	8.0	215.0	56671.0	67.0	0.238	0.893
19	0.0	2.0	1457.0	225972.0	415.0	0.222	0.995	0.018	7.0	342.0	56544.0	68.0	0.166	0.907
20	0.0	0.0	1132.0	226297.0	417.0	0.269	1.0	0.011	10.0	172.0	56714.0	65.0	0.274	0.867
21	0.0	1.0	840.0	226589.0	416.0	0.331	0.998	0.008	11.0	100.0	56786.0	64.0	0.39	0.853
22	0.0	1.0	2124.0	225305.0	416.0	0.164	0.998	0.075	10.0	350.0	56536.0	65.0	0.157	0.867
23	0.0	2.0	1457.0	225972.0	415.0	0.222	0.995	0.03	11.0	242.0	56644.0	64.0	0.209	0.853
24	0.0	5.0	2761.0	224668.0	412.0	0.13	0.988	0.297	6.0	2741.0	54145.0	69.0	0.025	0.92
25	0.0	3.0	2484.0	224945.0	414.0	0.143	0.993	0.025	10.0	199.0	56687.0	65.0	0.246	0.867
26	0.0	4.0	4867.0	222562.0	413.0	0.078	0.99	0.021	18.0	33.0	56853.0	57.0	0.633	0.76
27	0.0	8.0	4230.0	223199.0	409.0	0.088	0.981	0.053	9.0	1541.0	55345.0	66.0	0.041	0.88
28	0.0	9.0	5305.0	222124.0	408.0	0.071	0.978	0.026	9.0	398.0	56488.0	66.0	0.142	0.88
29	0.0	5.0	4846.0	222583.0	412.0	0.078	0.988	0.242	6.0	7883.0	49003.0	69.0	0.009	0.92
30	0.0	5.0	5193.0	222236.0	412.0	0.074	0.988	0.026	7.0	449.0	56437.0	68.0	0.132	0.907

Model Plot

View Model Plot