how to fix the incorrect dimension of numpy array

If you are using scikit-learn SVM classification algorithm, it expects 2D arrays of shape (n_samples, n_features) for the training dataset for a SVM fit function.

The dataset you are passing in is a 4D array, therefore you need to reshape the array into a 2D array.

Example:

from sklearn.model_selection import train_test_split
from sklearn.svm import SVC

# To apply a classifier, we need to flatten the image, to
# turn the data in a (samples, feature) matrix, 
# assuming data is numpy array of shape (17500, 32, 32, 3), convert to shape (17500, 3072).
n_samples = len(data)
data_reshape = data.reshape((n_samples, -1))

# Split data into train and test subsets
X_train, X_test, y_train, y_test = train_test_split(data_reshape, labels, 
                                                    test_size=0.2)
clf.fit(X_train,y_train)

The technique is called Dimensionality Reduction. Mapping data from high dimensional space into lower dimensions. The most commonly used technique is the Principal Component Analysis(PCA). You can learn about them through the following links :

• https://towardsdatascience.com/feature-selection-and-dimensionality-reduction-f488d1a035de

• https://www.quora.com/What-dimensionality-reduction-methods-would-you-recommend

  This link explains reduction with an example which has a dataset similar to yours : https://www.datacamp.com/community/tutorials/principal-component-analysis-in-python