ABSTRACT
This chapter digs deeper into techniques that are more practical in terms, not imposing the limitation that matrices must be square. It looks at how PCA can be used to create a covariance matrix and correlation matrix. Like Principal Component Analysis, Linear Discriminant Analysis (LDA) is used for dimension reduction. While PCA focuses on determining the principal component axes, those axes that maximize the variance of the data, Linear Discriminant Analysis also attempts to maximize the spread of the clusters or classes in the data. Some digital humanists focus on semantics and words that are regularly used or used in a particular context. Law enforcement agents have to deal with the issue of facial recognition on a daily basis. Similar ideas can be applied to handwriting samples and other image recognition problems.
