coranking package

Submodules

coranking.metrics module

coranking.metrics.LCMC(Q, min_k=1, max_k=None)

Compute the local continuity meta-criteria (LCMC) metric over a range of K values.

Parameters:

• Q – coranking matrix
• min_k – the lowest K value to compute. Default 1.
• max_k – the highest K value to compute. If None the range of values will be computer from min_k to n-1

Returns:

array of size min_k - max_k with the corresponding LCMC values.

coranking.metrics.continuity(Q, min_k=1, max_k=None)

Compute the continuity metric over a range of K values.

Parameters:

• Q – coranking matrix
• min_k – the lowest K value to compute. Default 1.
• max_k – the highest K value to compute. If None the range of values will be computer from min_k to n-1

Returns:

array of size min_k - max_k with the corresponding continuity values.

coranking.metrics.trustworthiness(Q, min_k=1, max_k=None)

Compute the trustwortiness metric over a range of K values.

Parameters:

• Q – coranking matrix
• min_k – the lowest K value to compute. Default 1.
• max_k – the highest K value to compute. If None the range of values will be computer from min_k to n-1

Returns:

array of size min_k - max_k with the corresponding trustworthiness values.

Module contents

coranking.coranking_matrix(high_data, low_data)

Generate a co-ranking matrix from two data frames of high and low dimensional data.

Parameters:

• high_data – DataFrame containing the higher dimensional data.
• low_data – DataFrame containing the lower dimensional data.

Returns:

the co-ranking matrix of the two data sets.