Preprocessing

MGP_AttTCN Framework

Get Started. It's Free
or sign up with your email address
Preprocessing by Mind Map: Preprocessing

1. data_preprocessing.extract_MIMIC_data.extract_labels.make_labels import make_labels

1.1. Generate sepsis onset time by hadm_id

1.2. Generate SOFA

2. data_preprocessing.extract_MIMIC_data.extract_features.make_data import MakeData

2.1. ICU data extraction

3. merge Data

3.1. dir: interim

3.1.1. - static_variables// vital variables // cases and controls

4. Step I Preprocessing

4.1. data_preprocessing.features_preprocessing.stepI_data_prep import DataPreprocessing

4.2. processed/full_static.csv

4.3. processed/onset_hours.csv

4.4. processed/full_labvitals_horizon_0_last.csv

5. Step II Split & Normalize

5.1. split into train/test/val and normalize

5.2. train/full_static.csv

5.3. train/full_labvitals.csv

6. Step III: GP I: Compact Transform

6.1. Transforms data for GP

6.2. GP_prep.pkl

7. Step IV: GP II: GPPreprocessing

7.1. GP_prep_v2.pkl

8. src/mains/main

8.1. Data Generator

8.1.1. load data (GP_prep_v2.pkl)

8.1.2. remove IDs

8.1.3. separate case and control data (train data)

8.1.3.1. balance the dataset to 50/50

8.1.4. list of patients present at horizon 6

8.1.4.1. late_case_patients

8.1.4.2. late_control_patients

8.2. Model

8.3. Trainer

8.3.1. Data

8.3.1.1. reshuffle

8.3.1.2. for every train epoch: for every batch

8.3.1.2.1. batch data = Y, T, ind_features, num_distinct_Y, X, num_distinct_X, static, labels, classes