LECA.prep.feature_overview

LECA.prep.feature_overview(data: DataFrame, objective_funcs: str | List[str], features: str | List[str], fig_size: Tuple[int, int] = (6, 4), save_loc: bool | str = False) → None

Outputs correlation/covariance plots for features along with feature importance plots for each objective function (uses sklearn.ensemble.RandomForestRegressor with default hyperparameters)

Parameters:

• data (DataFrame) – Dataframe of experimental measurements

• objective_funcs (Union[str, List[str]]) – Objective function or list of objective functions

• features (Union[str, List[str]]) – Feature or list of features

• fig_size (Tuple[int,int]) –

  Size of plots (follows matplotlib convention)

  By default (6,4).

• save_loc (Union[bool, str]) –

  Name to save plots (if desired), if False the plots will only be shown, not saved.

  Saving filename convention is:

  • save_loc + ‘corr.pdf’

  • save_loc + ‘cov.pdf’

  • save_loc + objective function name + ‘-feature-importance.pdf’

Return type:

None