“Predictive data analytics” is a new term coined in the context of machine learning, where such approaches tend to provide analytic predictions about voluminous data. Machine learning approaches are used to the huge amount of data and proposing predictions about it. This book covers machine learning for predictive data analytics.
The book comprises 11 chapters with an exhaustive bibliography of more than 100 references, a detailed index for referencing, and three appendices. The first chapter provides an introduction to machine learning for predictive data analytics. Chapters 2 and 3 focus on “data to insights to decisions,” and exploration using a motor insurance fraud case study, respectively. The next four chapters discuss information-, similarity-, probability-, and error-based learning methodologies using standard approaches like the ID3 algorithm, the nearest neighbor algorithm, the naive Bayes model, and multivariate linear regression with gradient descent. Chapter 8 details evaluation strategies using the misclassification rate on a hold-out test set as the standard approach, with categorical targets, prediction scores, multinomial targets, and continuous targets as the performance measures. Chapters 9 and 10 deal with customer churn and galaxy classification case studies. In the last chapter, the authors present the art of matching machine learning approaches to projects and data for predictive data analytics.
The appendices provide details about descriptive statistics and data visualization, an introduction to probability, and differentiation techniques for machine learning. There is also an online resources link as an additional feature to reading, writing code, and understanding the workings of machine learning approaches in predictive data analytics, which makes this book an interesting read. The intended audience includes undergraduates, postgraduates, practitioners, and professionals working in the predictive data analytics area.
More reviews about this item: Amazon