Summary

In this chapter, we learned about eigenvalues, eigenvectors, and orthogonal bases and how these concepts connect to form a basis for dimensionality reduction. We then learned about the two types of dimensionality reduction methods – feature elimination and feature extraction. We discussed the different steps of performing Principal Component Analysis which falls into the feature extraction category for dimensionality reduction. We used the implementation of PCA from scikit-learn to apply the algorithm to our dataset, where we reduced the features in our pizza dataset from 7 to 2 and visualized the data. We were able to easily tell that the nutrients present in the pizzas manufactured by different companies were different. Lastly, we applied PCA to the MNIST dataset and figured out that only 300 principal components were needed to capture 90% of the variance in the dataset, as compared to the 784 feature variables that we had originally, reducing the dimensionality by...