if *onehot {
data, cols = recodeOnehot(data, cols)
}
- pca, err := nlp.NewPCA(*components).FitTransform(array2matrix(rows, cols, data))
+ pca, err := nlp.NewPCA(*components).FitTransform(array2matrix(rows, cols, data).T())
if err != nil {
return 1
}
+ pca = pca.T()
rows, cols = pca.Dims()
out := make([]float64, rows*cols)
for i := 0; i < rows; i++ {