sica.mutualknn.MNN

class sica.mutualknn.MNN(k, metric, X, Y=None)

Given two arrays X and Y or a precomputed distance matrix computes the undirected adjacency matrix using the Mutual Nearest Neighbors method.

Parameters:

X2D array of shape (n_components_1 , n_features_1), 2D array of shape (n_components_1 , n_components_2)

if metric = “precomputed”.

Y2D array of shape (n_components_2 , n_features_2), optional

The default is None.

kint > 0

Parameter for the Mutual Nearest Neighbor method (i.e. number of neighbors that we consider).

metricstring

Metric for the computation of the adjacency matrix (e.g “pearson” , “spearman” or any metric accepted by scipy.spatial.distance.cdsit).

Attributes:

min_dist_float

minimal non-null distance between two different components.

max_dist_float

maximal non-null distance between two different components.

Notes

• In the case where the distance matrix is not precomputed, we compute the distance between each rows of X and Y.

• In the case X and Y are dataframes, we consider only the common columns of X and Y. Otherwise, we assume that the

columns are the same for X and Y.

Methods

adjacency_matrix(weighted)	Compute the undirected adjacency matrix with the Mutual Nearest Neighbors method (k neighbors)
compute_distance(X, Y, metric)	Compute the distance between each pair of rows of X and Y

adjacency_matrix(weighted)

Compute the undirected adjacency matrix with the Mutual Nearest Neighbors method (k neighbors)

Parameters:

weightedbool

If True each coefficient of the adjacency matrix is weighted by the associated distance,

otherwise the coefficients are 0 or 1.

Returns:

2D array of shape (n_components_1 , n_components_2)

static compute_distance(X, Y, metric)

Compute the distance between each pair of rows of X and Y

Parameters:

X2D array of shape (n_components_1 , n_features_1)

Y2D array of shape (n_components_2 , n_features_2)

metricstring

Returns:

2D array of shape (n_components_1 , n_components_2)