compshs.visualization package

Subpackages

Submodules

compshs.visualization.plot module

Created in 2025 @author: Simon Delarue <simon.delarue@telecom-paris.fr>

compshs.visualization.plot.plot_feature_selection(corpus, category, category_name, not_category_name, min_term_frequency)[source]

Plot feature selection in html using ScatterText library.

Parameters

  • corpus –

  • category –

  • category_name –

  • not_category_name –

  • min_term_frequency –

compshs.visualization.plot.plot_fixed_approaching(similarities: DataFrame, metric: str, keyword_order: list, keyword_colors: dict)[source]

Barchart for approaching-based semantic shift detection results for fixed mode.

Parameters

  • similarities (pd.DataFrame) – Semantic shift detection results.

  • metric (str) – Name of semantic shift detection metric to plot ({'sapp', 'asapp'}).

  • keyword_order (list) – Legend ordering.

  • keyword_colors (dict) – Dictionary with keywords as keys and corresponding colour as values (both str).

compshs.visualization.plot.plot_pyLDA(topic_modeler: TopicModeler, matrix: Union[csr_matrix, ndarray], counter: FrequencyCounter)[source]

Plot LDA using pyLDAvis library.

Parameters

  • topic_modeler (TopicModeler) – Fitted topic modeler.

  • matrix (sparse.csr_matrix, np.ndarray) – Document term matrix (n_documents, n_words).

  • counter (FrequencyCounter) – Frequency counter.

Return type

Visualization object.

compshs.visualization.plot.plot_sequential_approaching(similarities: DataFrame, metric: str = 'sapp')[source]

Heatmap for approaching-based semantic shift results in sequential mode.

Parameters

  • similarities (pd.DataFrame) – Semantic shift detection results.

  • metric (str) – Name of semantic shift detection metric to plot ({'sapp', 'asapp'}).

compshs.visualization.plot.plot_ssta(similarities: DataFrame)[source]

Plot SSTA.

Parameters

similarities (pd.DataFrame) – DataFrame of embedding similarities (such as SSTA.transform() output).

compshs.visualization.plot.plot_top_words(topic_modeler: TopicModeler, token_names: ndarray, k: int = 5, title: Optional[str] = None) → Figure[source]

Plot top \(k\) tokens in each modeled topic.

Parameters

  • topic_modeler (TopicModeler) – Fitted topic modeler.

  • token_names (np.ndarray) – Array of token names.

  • k (int) – Number of tokens displayed per topic (default = 5).

  • title (str) – Plot title. If None, the name of the model is used.

Return type

Figure

References

Scikit-learn documentation (see https://scikit-learn.org/stable/lite/lab/index.html).

Module contents

visualization module