应用预测建模(英文版)

本书是一部关于数据分 析的经典教材，聚焦预测建 模的实际应用，如如何进行 数据预处理、模型调优、预 测变量重要性度量、变量选 择等。读者可以从中学到许 多建模方法以及提高对许多 常用的、现代的有效模型的 认识，如线性回归、非线性 回归和分类模型，涉及树方 法、支持向量机等。书中还 涉及从数据预处理到建模再 到模型评估和选择的整个过 程，以及背后的统计思想， 涉及各种回归技术和分类技 术。

1 Introduction 1.1 Prediction Versus Interpretation 1.2 Key Ingredients of Predictive Models 1.3 Terminology 1.4 Example Data Sets and Typical Data Scenarios 1.5 Overview 1.6 Notation Part Ⅰ General Strategies 2 A Short Tour of the Predictive Modeling Process 2.1 Case Study: Predicting Fuel Economy 2.2 Themes 2.3 Summary 3 Data Pre-processing 3.1 Case Study: Cell Segmentation in High-Content Screening 3.2 Data Transformations for Individual Predictors 3.3 Data Transformations for Multiple Predictors 3.4 Dealing with Missing Values 3.5 Removing Predictors 3.6 Adding Predictors 3.7 Binning Predictors 3.8 Computing Exercises 4 Over-Fitting and Model Tuning 4.1 The Problem of Over-Fitting 4.2 Model Tuning 4.3 Data Splitting 4.4 Resampling Techniques 4.5 Case Study: Credit Scoring 4.6 Choosing Final Tuning Parameters 4.7 Data Splitting Recommendations 4.8 Choosing Between Models 4.9 Computing Exercises Part Ⅱ Regression Models 5 Measuring Performance in Regression Models 5.1 Quantitative Measures of Performance 5.2 The Variance-Bias Trade-off 5.3 Computing 6 Linear Regression and Its Cousins 6.1 Case Study: Quantitative Structure-Activity Relationshir Modeling 6.2 Linear Regression 6.3 Partial Least Squares 6.4 Penalized Models 6.5 Computing Exercises 7 Nonlinear Regression Models 7.1 Neural Networks 7.2 Multivariate Adaptive Regression Splines 7.3 Support Vector Machines 7.4 K-Nearest Neighbors