CODEWITHSUNDEEP
pythonregressionsparse-matrix
user1581390
user1581390

Reputation: 1998

SGDRegressor Won't Accept a List of Sparse Matrixes

I don't have the memory to convert the entire list of sparse matrices into a numpy 2d array, and then convert that to a sparse matrix.

The regressor WILL accept the following:

X = sparse.csr_matrix( my_2D_Numpy_Matrix )

It doesn't accept (this is just an example):

X = []
for i in range(my_2D_Numpy_Matrix.shape[0]):
    X.append(sparse.csr_matrix(my_2D_Numpy_Matrix[i,:]))

Upvotes: 0

Views: 92

Answers (1)

Ozan Bulut
Ozan Bulut

Reputation: 755

You can merge sparse matrices without converting them to numpy matrix using sparse.vstack

X = sparse_list[0]

for mat in sparse_list[1:]:
    X = sparse.vstack((X,mat))

Upvotes: 1

Related Questions