Feature selection is a fundamental step in Machine Learning (ML) that involves choosing some input data that would enhance the model performance. The model is able to run faster using lesser computational resources while giving reasonable results. Hence, feature selection as important as selection of a good model. In this chapter the aim is to analyze how the performance of different multiclass classification algorithms is affected on different features. The algorithms used are K-Nearest Neighbor (KNN), Support Vector Machines (SVM), Linear Discriminant Analysis (LDA), and Convolutional Neural Network (CNN) on the CIFAR-10 dataset. To obtain the new dataset with modified features, we use dimension reduction methods on the original dataset. The new dataset is at least 500x smaller, and we have noticed that in the best case scenarios reducing dimensions reduces the accuracy score only marginally. The SVM is the most consistent among the experimented models.