Fraud-Detection-MLOps/params.yaml at dev · Mark007-R/Fraud-Detection-MLOps · GitHub

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
data:
  raw_paysim_path: data/raw/paysim.csv
  raw_sparkov_path: data/raw/sparkov.csv
  combined_path: data/processed/combined_transactions.csv
  features_path: data/processed/features.csv
  x_test_path: data/processed/X_test.csv
  y_test_path: data/processed/y_test.csv
  model_path: models/fraud_model.pkl
  metrics_path: metrics/scores.json
  confusion_matrix_path: reports/confusion_matrix.svg
  roc_curve_path: reports/roc_curve.svg
  benchmark_path: metrics/fdb_benchmark.json

fdb:
  dataset_key: sparkov

combine:
  paysim_path: data/raw/paysim.csv
  sparkov_path: data/raw/sparkov.csv
  output_path: data/processed/combined_transactions.csv

preprocess:
  input: data/processed/combined_transactions.csv
  output: data/processed/features.csv

train:
  model_type: xgboost
  test_size: 0.2
  random_state: 42
  n_estimators: 200
  max_depth: 6
  learning_rate: 0.1
  scale_pos_weight: 50
  use_smote: false
  early_stopping_rounds: 0

evaluate:
  model_path: models/fraud_model.pkl
  metrics_path: metrics/scores.json
  reports_dir: reports/

benchmark:
  model_path: models/fraud_model.pkl
  output_path: metrics/fdb_benchmark.json
  baselines:
    AutoGluon: 0.952
    H2O AutoML: 0.947
    AutoSklearn: 0.931