About the Book

This book provides in introduction into data mining methods and models, including association rules, clustering, K-nearest neighbor, statistical inference, neural networks, linear and logistic regression, and multivariate analysis. It presents a unified approach based on CRISP methodology (involves Strategic Risk Assessment based on Organizational Modelling).

About the Author

Daniel T. Larose received his PhD in statistics from the University of Connecticut. Currently he is an associate professor of statistics in the Department of Mathematical Sciences, and Director of Data Mining@CCSU, at Central Connecticut State University. He has also worked in the areas of biostatistics, statistics, and data management at Bristol-Myers Squibb Pharmaceutical Research Center, Wesleyan University, and United Technologies Corporation. He was written and contributed work to two books, and published 9 articles, on the topic of data mining and statistics.

Table of Contents:
Preface Dimension Reduction Methods · Need for Dimension Reduction in Data Mining · Principal Components Analysis · Factor Analysis · User-Defined Composites Regression Modeling · Example of Simple Linear Regression · Least-Squares Estimates · Coefficient or Determination · Correlation Coefficient · The ANOVA Table · Outliers, High Leverage Points, and Influential Observations · The Regression Model · Inference in Regression · Verifying the Regression Assumptions · An Example: The Baseball Data Set · An Example: The California Data Set · Transformations to Achieve Linearity Multiple Regression and Model Building · An Example of Multiple Regression · The Multiple Regression Model · Inference in Multiple Regression · Regression with Categorical Predictors · Multicollinearity · Variable Selection Methods · An Application of Variable Selection Methods · Mallows’ C p Statistic · Variable Selection Criteria · Using the Principal Components as Predictors in Multiple Regression Logistic Regression · A Simple Example of Logistic Regression · Maximum Likelihood Estimation · Interpreting Logistic Regression Output · Inference: Are the Predictors Significant? · Interpreting the Logistic Regression Model · Interpreting a Logistic Regression Model for a Dichotomous Predictor · Interpreting a Logistic Regression Model for a Polychotomous Predictor · Interpreting a Logistic Regression Model for a Continuous Predictor · The Assumption of Linearity · The Zero-Cell Problem · Multiple Logistic Regression · Introducing Higher Order terms to Handle Non-Linearity · Validating the Logistic Regression Model · WEKA: Hands-On Analysis Using Logistic Regression Naïve Bayes and Bayesian Networks · The Bayesian Approach · The Maximum a Posteriori (MAP) Classification · The Posterior Odds Ratio · Balancing the Data · Naïve Bayes Classification · Numeric Predictors for Naïve Bayes Classification · WEKA: Hands-On Analysis Using Naïve Bayes · Bayesian Belief Networks · Using the Bayesian Network to Find Probabilities · WEKA: Hands-On Analysis Using Bayes Net Genetic Algorithms · Introduction to Genetic Algorithms · The Basic Framework of a Genetic Algorithm · A Simple Example of Genetic Algorithms at Work · Modifications and Enhancements: Selection · Modifications and enhancements: Crossover · Genetic Algorithms for Real-Valued Variables · Using Genetic Algorithms to Train a Neural Network · WEKA: Hands-On Analysis Using Genetic Algorithms Case Study: Modeling Response to Direct-Mail Marketing · The Cross-Industry Standard Process for Data Mining: CRISP-DM · Business Understanding Phase · Data Understanding and Data Preparation Phases · The Modeling Phase and the Evaluation Phase Index