MobileRabbitHoleProcessML

Preprocessing & Feature Generation

featureGenerationGroup.py

In its main method one can try the feature extraction for a single user

• takes Miriam's non-grouped sessions (RabbitHoleProcess\data\dataframes\sessions_with_features\*.pickle) as basis
• sessions are clustered into session-groups (methods imported from grouped_sessions.py)

This file implements the methods that actually create features, however its main method is just for test and dev purposes

(old: featureGeneration.py - the stuff Miriam did)

preprocessingGroup.py

Actually performs the preprocessing (if desired) for all users.

TODO input output?

ML_classification.py

TODO

Data

(with exemplary user AN23GE)

Miriam's features of non-grouped sessions

RabbitHoleProcess\\data\\dataframes\\sessions_with_features\\AN23GE.pickle One line per unlock/lock session, one column per generated feature. Includes columns for the ESM responses

All Raw Sensing Data

C:\\Users\\florianb\\Downloads\\AN23GE.pickle <- one example D:\usersorted_logs_preprocessed\\AN23GE.pickle <- all One line per sensing event log (further data on Flo's external HDD due to lack of storage)