So I have two matrix, W and X.
print(W) array([ 5.76951515, 19. ]) print(X) array([[ 1., 5.], [ 1., 6.], [ 1., 7.], [ 1., 8.], [ 1., 9.], [ 1., 10.], [ 1., 11.], [ 1., 12.], [ 1., 13.], [ 1., 14.]]) and I'd like multiply both matrix W and X, varying the value of W[1] for each i iterations, like this.
for i in range(10): W[1] = i yP_ = W @ X.T ecm = np.mean((Y - yP_ ) ** 2) plt.plot(W[1], ecm, 'o') plt.show() is there any way to avoid that for?
Try making W shape (10,2), and keeping the range 0-9 in the second column. Then the rows of the product W @ X.T are the iterations of your current for-loop.
W2 = np.full((10,2), W[0]) W2[:,1] = np.arange(10) W2 # array([[5.76951515, 0. ], # [5.76951515, 1. ], # ... # [5.76951515, 9. ]]) So you can do
ecm = np.mean((Y - W2 @ X.T)**2, axis=1) # average across columns plt.plot(W2[:,1], ecm, 'o') You can start by generating the modified W array, and then apply the matrix product just as you where:
N=10 W_ = np.c_[[W[0]]*N, np.arange(N)] yP_ = [email protected] Quick check:
yP_ = [] for i in range(N): W[1] = i yP_.append(W @ X.T) np.allclose(np.array(yP_), [email protected]) # True 
for? Performance?