pmc_maqc-ml/params.yaml at develop · babelomics/pmc_maqc-ml · 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
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
# -- General params --
raw_data_folder: raw/20250403/
outcome_data_folder: raw/outcome_stage_1-5/
repo_data_folder: ./data/
dataset: LC
random_state: 42
# -- Master feature generation params --
master_params:
  output_folder: features/master/
  features:
    numeric:
    - Age_at_Diagnosis
    categorical:
    - GENDER
    - RACE
    - PATIENT_STATE_CODE
    - PATIENT_REGION
    - STAGE_AT_DIAGNOSIS
    - CANCER_TYPE
    - STAGE
# -- feature generation params --
feature_params:
  # -- diag --
  diag_params:
    apply_func: # Function and its params to apply to process features
      # # - Initial counting icd codes approach
      # diag_count_featurization:
      #   map_file: ./src/DX_CD_to_broad_icd10.yaml
      #   mapped_column_name: broad_ICD10
      #   col_to_count: broad_ICD10
      # - llm embeddings approach
      diag_medical_embeddings:
        col_to_embed: DESCRIPTION
        model_name: dmis-lab/biobert-base-cased-v1.2
        max_length: 128
        batch_size: 16
  # -- Procdr --
  procdr_params:
    apply_func: # Function and its params to apply to process features
      # procdr_count:
      #   col_to_count: ID
  # -- drug --
  drug_params:
    apply_func:
  # -- moltest --
  moltest_params:
    apply_func:
  # -- test --
  test_params:
    apply_func:
# -- Train parametrization --
train_params:
  n_repeats: 33
  n_splits: 3
  kbest: 100
  # Here we define the file we will use for each table
  features_to_use:
    master: complete
    diag: care
    procdr:
    drug:
    moltest:
    test:
  transformers_to_use: # Define the transformers.py transformers to use for each file
    master:
      num: # Apply the 'num' transformer to this column
      - Age_at_Diagnosis
      cat: # Apply the 'cat' transformer to this column
      - GENDER
      - RACE
      - PATIENT_REGION
      - PATIENT_STATE_CODE
      - STAGE_AT_DIAGNOSIS
      - CANCER_TYPE
      - STAGE
    diag: diag
    procdr: procdr
  model:
    # -- Logistic Regression CV
    name: lrcv
    params:
      max_iter: 1000
      scoring: matthews_score
      class_weight: balanced
    # # -- EBM
    # name: ebm
    # params:
    #   n_jobs: 50
# -- Predict params --
predict_params:
  metric_for_best_model: test_matthews_score