Machine Learning and Time Series Models for Consumer Behavior Prediction in the Data Analytics Paradigm

Abstract

In the big data paradigm, everyday consumers produce an infinite amount of data with their decisions. These data are available to companies, researchers, and institutions, and it is crucial for them to process it correctly in order to understand consumer behavior. This chapter presents the core time series and machine learning models to analyze and process consumer behavior data in order to convert raw data into useful and informative analyses, predictions, and forecasts. To do that, it introduces, explains, applies, and analyzes the results of ARIMA models, regression analysis, artificial neural networks models, machine learning decision trees, bootstrap methods, and much more. Every technique is illustrated through the R programming language, and the R code is provided through the text in order to ensure replicability and serve as a hands-on manual.