mvpa2.measures.corrcoef.pearson_correlation

mvpa2.measures.corrcoef.pearson_correlation(x, y=None)

Computes pearson correlations on matrices

Parameters:

x: np.ndarray or Dataset

PxM array

y: np.ndarray or Dataset or None (the default).

PxN array. If None, then y=x.

Returns:

c: np.ndarray

MxN array with c[i,j]=r(x[:,i],y[:,j])

Notes

Unlike numpy. this function behaves like matlab’s ‘corr’ function. Its numerical precision is slightly lower than numpy’s correlate function. Unlike scipy’s ‘pearsonr’ function it does not return p values. TODO integrate with CorrCoef