musicaiz.eval.model_features_violinplot¶
- musicaiz.eval.model_features_violinplot(measures_dist1: Dict[str, Dict[str, float]], measures_dist2: Dict[str, Dict[str, float]], label1: str = '1', label2: str = '2', measure: str = 'all', show: bool = True)[source]¶
Plots the distances distributions of the input features as a violin plot as the Fig. 4 of the paper by Yang and Lerch. Note that we normalize the distances vectors to plot them.
- Parameters
- measures_dist1: Dict[str, Dict[str, float]]
the dict with the cross-validated distances of each sample in the database (intra or inter) for each feature.
- measures_dist2: Dict[str, Dict[str, float]]
the dict with the cross-validated distances of each sample in the database (intra or inter) for each feature.
- label1: str
the label or name of the dataset of model which corresponds to the measures_dict1, for showing it in the legend.
- label2: str
the label or name of the dataset of model which corresponds to the measures_dict2, for showing it in the legend.