python code error. ValueError: y should be a 1d array, got an array of shape (56000, 10) instead

Question

import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.linear_model import LogisticRegression
logistic_regression = LogisticRegression(random_state = 10)
logistic_regression.fit(X_train, y_train)
y_pred_logistic_regression = logistic_regression.predict(X_test)
print(y_pred_logistic_regression.shape)

ValueError: y should be a 1d array, got an array of shape (56000, 10)instead

Answer 1

Your y_train array is out of shape, assuming you are working with one hot encoded classes, the solution would be to convert it into a class index array:

logistic_regression.fit(X_train, y_train.argmax(axis=1))

example:

>> y_train
array([[0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],
       [1., 0., 0., 0., 0., 0., 0., 0., 0., 0.]])
# after conversion:
>> y_train.argmax(axis=1)
array([2, 3, 4, 0])