Predictive Analytics for Business using R

The fields of mathematical statistics, statistical graphics, computer science and operations research have created the rich set of methods now called Analytics. Often analytics is characterized along three poles: descriptive analytics (what do data tell us), predictive analytics (what can be forecast based on the data, and with what certainty), and prescriptive analytics (how can the data inform changes to improve system performance).

This book focuses on the second pole, predictive analytics. The areas of predicting a number, a class, and dynamic behavior are distinct, with different methods. This text has three parts based on these areas. Topics in predicting a number include simple and multiple linear regression, transformation of variables, analysis of observational data via cross-validation, the generalized linear model, designed experiments, and Gaussian process and neural network regression. Classification methods include neural networks, logistic regression, k-nearest neighbor, and linear discriminant analysis. Methods for predicting dynamic behavior include trend analysis, time series analysis and discrete-event dynamic simulation.

Characterizing prediction uncertainty is a key focus of this text. The text provides analytic methods appropriate to each area, with an explicit process for applying such methods. Case data with corresponding R code are used to illustrate each method.

Predictive Analytics for Business using R is designed for a hybrid class structure. Class sessions can be a blend of lecture format and flipped classroom case analyses. In a two-meetings-per-week fifteen-week structure, one day per week would be devoted to explaining methodology and presenting a case study, with the second day focused on coaching. Given the case structure, the text does not contain homework problems. Instead, at the end of each chapter there are links to cases posted online.

Contents:

• Getting Started:

  • Introduction
  • The R Programming Environment

• Predicting a Number:

  • Linear Regression
  • Regression Under the Data-Adaptive Research Paradigm
  • Nontraditional Regression (Machine Learning)

• Predicting a Class:

  • Classification for Predictive Analyticss

• Predicting Dynamic Behavior:

  • Trendline and Time Series Analysis for Forecasting
  • Simulating Random Values
  • Building Discrete-Event Simulation Models
  • Prediction with Discrete-Event Simulation Models

Readership: Advanced undergraduates, Master's students in business, accounting, engineering, statistics or computer science, as well as practitioners in business analytics, business consulting, industrial engineering and operations research.
Key Features:

• Business analytics is a hot topic in business and business schools, and Predictive analytics is the central topic. Predictive analytics process is also developed hand in hand with all required R code
• To quote Jian Liu: "[The] author presents complex mathematical concepts in a clear and accessible manner, ensuring that readers with varying levels of expertise can grasp the fundamentals..."
• This book is appropriate for one-semester professional masters or senior undergraduates' courses in supply chain, finance or actuarial science
• This is the Only predictive analytics text to include discrete event simulation and topics are organized intuitively into areas like 'predicting a number', 'predicting a class', and 'predicting dynamic behavior'
• The book has cases and grading rubrics for each topic

Dettagli