sourcecode/scoring/pflip_model.py [804:813]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    if len(noteInfo) == 0:
      return
    logger.info(f"noteInfo summary: {get_df_fingerprint(noteInfo, [c.noteIdKey])}")
    trainDataFrame, validationDataFrame = train_test_split(noteInfo, train_size=self._trainSize)
    logger.info(f"pflip training data size: {len(trainDataFrame)}")
    logger.info(f"trainDataFrame summary: {get_df_fingerprint(trainDataFrame, [c.noteIdKey])}")
    logger.info(f"pflip validation data size: {len(validationDataFrame)}")
    logger.info(
      f"validationDataFrame summary: {get_df_fingerprint(validationDataFrame, [c.noteIdKey])}"
    )
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



sourcecode/scoring/pflip_plus_model.py [1586:1597]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    if len(noteInfo) == 0:
      return
    logger.info(f"noteInfo summary: {get_df_fingerprint(noteInfo, [c.noteIdKey])}")
    # Dividing training data temporally provides a more accurate measurement, but would also
    # require excluding the newest data from training.
    trainDataFrame, validationDataFrame = train_test_split(noteInfo, train_size=self._trainSize)
    logger.info(f"pflip training data size: {len(trainDataFrame)}")
    logger.info(f"trainDataFrame summary: {get_df_fingerprint(trainDataFrame, [c.noteIdKey])}")
    logger.info(f"pflip validation data size: {len(validationDataFrame)}")
    logger.info(
      f"validationDataFrame summary: {get_df_fingerprint(validationDataFrame, [c.noteIdKey])}"
    )
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -