stereo.core.StPipeline.gaussian_smooth#

StPipeline.gaussian_smooth(n_neighbors=10, smooth_threshold=90, pca_res_key='pca', n_jobs=-1, inplace=True)[source]#

Smooth the express matrix by the algorithm of Gaussian smoothing [Shen22].

Parameters:

  • n_neighbors (int) – the number of the nearest points to search. Too high value may cause overfitting, and too low value may cause porr smoothing effect.

  • smooth_threshold (int) – the threshold that indicates Gaussian variance with a value between 20 and 100. Also too high value may cause overfitting, and low value may cause poor smoothing effect.

  • pca_res_key (str) – the key of PCA to get targeted result from self.result.

  • n_jobs (int) – the number of parallel jobs to run, if -1, all CPUs will be used.

  • inplace (bool) – whether to inplace the previous express matrix or get a new StereoExpData object with the new express matrix. # noqa

Returns:

An object of StereoExpData with the express matrix processed by Gaussian smooting.