Functions
def	get_unique_features (features, threshold=0.025, chunksize=10, _bottom=False)

def	spearmanr (x, y)

Function Documentation

◆ get_unique_features()

def MDN.transformers.utils.get_unique_features	(	features,
		threshold = `0.025`,
		chunksize = `10`,
		_bottom = `False`
	)

Mask out features which are near duplicates of some other 
    feature(s), based on correlation. The threshold parameter
    determines how similar features can be to keep them.
    Roughly based on https://stackoverflow.com/a/66326102

Definition at line 7 of file utils.py.

◆ spearmanr()

def MDN.transformers.utils.spearmanr	(	x,
		y
	)

Scipy's provided spearmanr is too slow for large matrices, and so we instead use a custom implementation.
    Source: https://stackoverflow.com/questions/52371329/fast-spearman-correlation-between-two-pandas-dataframes/59072032#59072032

Definition at line 58 of file utils.py.

Functions

Function Documentation

◆ get_unique_features()

◆ spearmanr()