Status

Functional code

structure

.
├── data                   # datasets
├── src                    # source code 
│   ├── configs            # configs files, to edit for new experiments
│   ├── features           # folder to process the features for the classification pipeline ***
│   ├── ml                 # ML pipeline *** 
│   ├── ml                 # Pattern mining ***
├── notebooks              # notebooks to plot the results
├── results                # automatically generated results folder
└── README.md              # Hello World

Run

Classification

• add your dataset in the data folder
• add your own sequencer in src/features/sequencers, such that you can use it in pipeline_maker.py
• edit pipeline_maker.py to add your dataset as an option when using the config file
• add your own models (if needed) in ml/models/ based on the model superclass such that it can be used in xval_maker.py

run python script_classification.py --seeds

Pattern mining

• add your dataset in the data folder

• add your own sequencer in src/pattern_mining/features/sequencers, such that you can use it in pm_pipeline.py

• edit data_pipeline to add your dataset as an option when using the config file

• add your own models (if needed) in ml/models/ based on the model superclass such that it can be used in pm_pipeline

• all files with the word "config" in also need to be edited to make the pipeline run !

run pyton script_patternmining.py --mining --sequences

Files information

exampledataset_config.yaml This file is used to decide what across which demographic groups the differential pattern mining algorithm need to be, and gives information about where to find the dataset, and what demographic attributes are available.

pipeline_maker Used to load the sequences