2.4.7 Segment Filters

Segment filters are used to filter out segments based on statistical properties before calling feature generation and classification.

Segment Energy Threshold Filter

Takes the absolute value of each point and compares it with the threshold. If no values are above the threshold the segment is filtered.

Parameters

  • input_data (_type_) – _description_

  • input_column (_type_) – _description_

  • group_columns (_type_) – _description_

  • threshold (int , optional) – _description_. Defaults to 0.

  • backoff (int , optional) – _description_. Defaults to 0.

  • delay (int , optional) – _description_. Defaults to 0.

  • disable_train (bool , optional) – _description_. Defaults to False.

Returns

Filtered segments

Segment Filter MSE

Filters out groups that does not pass the MSE threshold.

Parameters

  • input_column (str) – the name of the column to use for filtering

  • MSE_target (float) – filter target value (default is -1.0)

  • MSE_threshold (float) – filter threshold value (default is 0.01)

Returns

The filtered dataframe.

Return type

DataFrame

Segment Filter Threshold

Filters out segments that have values greater or less than the specified threshold.

Parameters

  • input_column (str) – the name of the column to use for filtering

  • threshold (int16) – threshold value to filter against

  • comparison (int) – 0 for less than, 1 for greater than

Returns

The filtered dataframe.

Return type

DataFrame