预期的2D阵列，改为1D阵列

Question

I am getting this error even though after much debugging, the array seems to be a 2D array:

    reg = linear_model.SGDClassifier()
    X = []
    Y = []
    with open('data/legitimate.txt', 'r', encoding='utf8') as f:
        for line in f:
            X.append(get_heuristics(line))
            Y.append(0)

    with open('data/bad.txt', 'r', encoding='utf8') as f:
        for line in f:
            X.append(get_heuristics(line))
            Y.append(1)

    X = np.array(X)
    Y = np.array(Y)
    reg.fit(X, Y)

Where get_heuristics() returns a 1D array. Any reason why this could be happening? Thanks

Answer 1

The X array is a 2D array. As you correctly say getheuristics() returns a 1D array and appending a 1D array will make a 2D array. However the error "Expected 2D array, got 1D array instead" will be displayed because Y is only 1D. After all, you are simply appending a single digit each time, hence a 1D array. Scikit-learn's fit requires two 2D arrays. As I'm sure the console will have suggested, you might want to try:

Y = Y.reshape(-1, 1)

before you call reg.fit(X,Y) .