scmags.ScMags.knn_classifier

ScMags.knn_classifier(test_ratio=0.3, nof_neighbors=3, log_norm=True, conf_normalize='true', rand_state=None, main_title='Confusion Matrix', figsize=None, cmap=<matplotlib.colors.LinearSegmentedColormap object>, plot_name='scmags_knn_result', save_plot=False, plot_dpi=300)

This function performs k-NN classification with selected markers. Visualizes the results with a normalized confusion matrix.

Parameters

• test_ratio (Optional[float]) – Test data rate. Should be between 0-1.

• nof_neighbors (Optional[int]) – Number of neighbors to use.

• log_norm (Optional[bool]) – Performs log normalization before k-NN.

• conf_normalize (Optional[Literal[‘true’, ‘pred’, ‘all’]]) – Normalizes confusion matrix over the true (rows), predicted (columns) conditions or all the population. If None, confusion matrix will not be normalized.

• rand_state (Optional[int]) – Controls the shuffling applied to the data before applying the split.

• main_title (Optional[str]) – Main title for confusion matrix.

• figsize (Optional[Tuple[int, int]]) – Figure size. If not given its own settings.

• cmap (Optional[str], optional) – A Colormap instance or registered colormap name. (matplotlib cmap)

• plot_name (Optional[str]) – Name of plot to save

• save_plot (bool) – If True, the plot is saved in the working directory.

• plot_dpi (int) – The dpi value of the figure to be saved

Examples

>>> import scmags as mg
>>> li = mg.datasets.li()
>>> li.filter_genes()
>>> li.sel_clust_marker()
>>> li.knn_classifier()