Stationary Signal, Autocorrelation, and Linear and Discriminant Analysis

Abstract

The effect of the divergent nature of signals under a process enables the computer system to interpret the signals, i.e., either analog or digital in an appropriate way so as to perform as per the given instructions. The process of signal interpretation enables the machine to perform a wide range of applications that are signal specific in nature. The process of autocorrelation in machine learning (ML) is the mathematical representation of the degree of similarity among the provided time series and the lagged version of itself spread over the successive time intervals. The concept of autocorrelation process in ML has specifically been designed in order to support the process of similarity measurement among the lagged values and the actual value within the provided time series. Linear discriminant analysis (LDA), normal discriminant analysis (NDA), or discriminant function analysis, a generalized version of Fisher’s linear discriminant, is a method used in statistics, pattern recognition, and machine learning to find a linear combination of features which can characterize or discriminate among two or more classes of objects or events. The resultant combination may be further utilized as a linear classifier or for dimensionality reduction before later classification. The chapter discusses in detail the role of stationary signal, autocorrelation, and linear and discriminant analysis in machine learning to identify various patterns and distinct features in machines.