xplore(r) - application guide

ISBN-13: 9783540675457 / Angielski / Miękka / 2000 / 525 str.

Wolfgang Hardle; W. Hardle; W. Hdrdle

Opis

Most statistical applications involve computational work with data stored on a computer. The mechanics of interaction with the data is a function of the sta- tistical computing environment. This application guide is intended for slightly experienced statisticians in computer-aided data analysis who desire to learn advanced applications in various fields of statistics. The prerequisities for XploRe-the statistic computing environment-are an introductory course in statistics or mathematics. This book is designed as an e-book which means that the text contained in here is also available as an integrated document in HTML and PDF format. The reader of this application guide should therefore be familiar with the basics of Acrobat Reader and of HTML browsers in order to profit from direct computing possibilities within this document. The quantlets presented here may be used together with the academic edi- tion of XploRe (http: //www.i-xplore.de) or via the XploRe Quantlet Client (XQC) on http: //www.xplore-stat.de. The book comes together with a CD- Rom that contains the XploRe Quantlet Server (XQS) and the full Auto Pilot Support System (APSS). With this e-book bundle one may directly try the application without being dependent on a specific software version. The quantlets described in the book can be accessed via the links included All executable quantlets are denoted by the symbol . Some in the text.

Spis treści

I Regression Models.- 1 Quantile Regression.- 1.1 Introduction.- 1.2 Quantile Regression.- 1.2.1 Definitions.- 1.2.2 Computation.- 1.3 Essential Properties.- 1.3.1 Equivariance.- 1.3.2 Invariance to Transformations.- 1.3.3 Robustness.- 1.4 Inference.- 1.4.1 Main Asymptotic Results.- 1.4.2 Wald Test.- 1.4.3 Rank Tests.- 1.5 Description of Quantlets.- 1.5.1 Quantlet rqfit.- 1.5.2 Quantlet rrstest.- 2 Least Trimmed Squares.- 2.1 Robust Regression.- 2.1.1 Introduction.- 2.1.2 High Breakdown point Estimators.- 2.2 Least Trimmed Squares.- 2.2.1 Definition.- 2.2.2 Computation.- 2.3 Supplementary Remarks.- 2.3.1 Choice of the Trimming Constant.- 2.3.2 LTS as a Diagnostic Tool.- 2.3.3 High Subsample Sensitivity.- 3 Errors-in-Variables Models.- 3.1 Linear EIV Models.- 3.1.1 A Single Explanatory Variable.- 3.1.2 Vector of Explanatory Variables.- 3.2 Nonlinear EIV Models.- 3.2.1 Regression Calibration.- 3.2.2 Simulation Extrapolation.- 3.3 Partially Linear EIV Models.- 3.3.1 The Variance of Error Known.- 3.3.2 The Variance of Error Unknown.- 3.3.3 XploRe Calculation and Practical Data.- 4 Simultaneuos-Equations Models.- 4.1 Introduction.- 4.2 Estimation.- 4.2.1 Identification.- 4.2.2 Some Notation.- 4.2.3 Two-Stage Least Squares.- 4.2.4 Three-Stage Least Squares.- 4.2.5 Computation.- 4.3 Application: Money-Demand.- 5 Hazard Regression.- 5.1 Data Structure.- 5.2 Kaplan-Meier Estimates.- 5.3 The Cox Proportional Hazards Model.- 5.3.1 Estimating the Regression Coefficients.- 5.3.2 Estimating the Hazard and Survival Functions.- 5.3.3 Hypothesis Testing.- 5.3.4 Example: Length of Stay in Nursing Homes.- 6 Generalized Partial Linear Models.- 6.1 Estimating GPLMs.- 6.1.1 Models.- 6.1.2 Semiparametric Likelihood.- 6.2 Data Preparation.- 6.2.1 General.- 6.2.2 Example.- 6.3 Computing GPLM Estimates.- 6.3.1 Estimation.- 6.3.2 Estimation in Expert Mode.- 6.4 Options.- 6.4.1 Setting Options.- 6.4.2 Grid and Starting Values.- 6.4.3 Weights and Offsets.- 6.4.4 Control Parameters.- 6.4.5 Model Parameters.- 6.4.6 Specification Test.- 6.4.7 Output Modification.- 6.5 Statistical Evaluation and Presentation.- 6.5.1 Statistical Characteristics.- 6.5.2 Output Display.- 6.5.3 Model selection.- 7 Generalized Additive Models.- 7.1 Brief Theory.- 7.1.1 Models.- 7.1.2 Marginal Integration.- 7.1.3 Backfitting.- 7.1.4 Orthogonal Series.- 7.2 Data Preparation.- 7.3 Noninteractive Quantlets for Estimation.- 7.3.1 Estimating an AM.- 7.3.2 Estimating an APLM.- 7.3.3 Estimating an AM and APLM.- 7.3.4 Estimating a GAM.- 7.3.5 Estimating a GAPLM.- 7.3.6 Estimating Bivariate Marginal Influence.- 7.3.7 Estimating an AM with Interaction Terms.- 7.3.8 Estimating an AM Using Marginal Integration.- 7.4 Interactive Quantlet GAMFIT.- 7.5 How to Append Optional Parameters.- 7.6 Noninteractive Quantlets for Testing.- 7.6.1 Component Analysis in APL Models.- 7.6.2 Testing for Interaction.- 7.6.3 Testing for Interaction.- 7.7 Odds and Ends.- 7.7.1 Special Properties of GAM Quantlib Quantlets.- 7.7.2 Estimation on Principal Component by PCAD.- 7.8 Application for Real Data.- II Data Exploration.- 8 Growth Regression and Counterfactual Income Dynamics.- 8.1 A Linear Convergence Equation.- 8.2 Counterfactual Income Dynamics.- 8.2.1 Sources of the Growth Differential With Respect to a Hypothetical Average Economy.- 8.2.2 Univariate Kernel Density Estimation and Bandwidth Selection.- 8.2.3 Multivariate Kernel Density Estimation.- 9 Cluster Analysis.- 9.1 Introduction.- 9.1.1 Distance Measures.- 9.1.2 Similarity of Objects.- 9.2 Hierarchical Clustering.- 9.2.1 Agglomerative Hierarchical Methods.- 9.2.2 Divisive Hierarchical Methods.- 9.3 Nonhierarchical Clustering.- 9.3.1 K-means Method.- 9.3.2 Adaptive K-means Method.- 9.3.3 Hard C-means Method.- 9.3.4 Fuzzy C-means Method.- 10 Classification and Regression Trees.- 10.1 Growing the Tree.- 10.2 Pruning the Tree.- 10.3 Selecting the Final Tree.- 10.4 Plotting the Result of CART.- 10.5 Examples.- 10.5.1 Simulated Example.- 10.5.2 Boston Housing Data.- 10.5.3 Density Estimation.- 11 DPLS: Partial Least Squares Program.- 11.1 Introduction.- 11.2 Theoretical Background.- 11.2.1 The Dynamic Path Model DPLS.- 11.2.2 PLS Estimation with Dynamic Inner Approximation.- 11.2.3 Prediction and Goodness of Fit.- 11.3 Estimating a DPLS-Model.- 11.3.1 The Computer Program DPLS.- 11.3.2 Creating design-matrices.- 11.3.3 Estimating with DPLS.- 11.3.4 Measuring the Forecasting Validity.- 11.4 Example: A Model for German Share Prices.- 11.4.1 The General Path Model.- 11.4.2 Manifest Variables and Sources of Data.- 11.4.3 Empirical Results.- 12 Uncovered Interest Parity.- 12.1 The Uncovered Interest Parity.- 12.2 The Data.- 12.3 A Fixed Effects Model.- 12.4 A Dynamic Panel Data Model.- 12.5 Unit Root Tests for Panel Data.- 12.6 Conclusions.- 12.7 Macro Data.- 13 Correspondence Analysis.- 13.1 Introduction.- 13.1.1 Singular Value Decomposition.- 13.1.2 Coordinates of Factors.- 13.2 XploRe Implementation.- 13.3 Example: Eye-Hair.- 13.3.1 Description of Data.- 13.3.2 Calling the Quantlet.- 13.3.3 Documentation of Results.- 13.3.4 Eigenvalues.- 13.3.5 Contributions.- 13.3.6 Biplots.- 13.3.7 Brief Remark.- 13.4 Example: Media.- 13.4.1 Description of the Data Set.- 13.4.2 Calling the Quantlet.- 13.4.3 Brief Interpretation.- III Dynamic Statistical Systems.- 14 Long-Memory Analysis.- 14.1 Introduction.- 14.2 Model Indepependent Tests for 1(0) against 1(d).- 14.2.1 Robust Rescaled Range Statistic.- 14.2.2 The KPSS Statistic.- 14.2.3 The Rescaled Variance V/S Statistic.- 14.2.4 Nonpaxametric Test for 1(0).- 14.3 Semiparametric Estimators in the Spectral Domain.- 14.3.1 Log-periodogram Regression.- 14.3.2 Semiparametric Gaussian Estimator.- 15 ExploRing Persistence in Financial Time Series.- 15.1 Introduction.- 15.2 Hurst and Fractional Integration.- 15.2.1 Hurst Constant.- 15.2.2 Fractional Integration.- 15.3 Tests for 1(0) against fractional alternatives.- 15.4 Semiparametric estimation of difference parameter d.- 15.5 Exploiting the Data.- 15.5.1 Typical Spectral Shape.- 15.5.2 Typical Distribution: Mean, Variance, Skewness and Kur-tosis.- 15.6 The Data.- 15.7 The Quantlets.- 15.8 The Results.- 15.8.1 Equities.- 15.8.2 Exchange.- 15.9 Practical Considerations.- 15.9.1 Risk and Volatility.- 15.9.2 Estimating and Forecasting of Asset Prices.- 15.9.3 Portfolio Allocation Strategy.- 15.9.4 Diversification and Fractional Cointegration.- 15.9.5 MMAR and FIGARCH.- 15.10Conclusion.- 16 Flexible Time Series Analysis.- 16.1 Nonlinear Autoregressive Models of Order One.- 16.1.1 Estimation of the Conditional Mean.- 16.1.2 Bandwidth Selection.- 16.1.3 Diagnostics.- 16.1.4 Confidence Intervals.- 16.1.5 Derivative Estimation.- 16.2 Nonlinear Autoregressive Models of Higher Order.- 16.2.1 Estimation of the Conditional Mean.- 16.2.2 Bandwidth and Lag Selection.- 16.2.3 Plotting and Diagnostics.- 16.2.4 Estimation of the Conditional Volatility.- 17 Multiple Time Series Analysis.- 17.1 Getting Started.- 17.1.1 Data Preparation.- 17.1.2 Starting multi.- 17.2 Preliminary Analysis.- 17.2.1 Plotting the Data.- 17.2.2 Data Transformation.- 17.3 Specifying a VAR Model.- 17.3.1 Process Order.- 17.3.2 Model Estimation.- 17.3.3 Model Validation.- 17.4 Structural Analysis.- 17.4.1 Impulse Response Analysis.- 17.4.2 Confidence Intervals for Impulse Responses.- 18 Robust Kalman Filtering.- 18.1 State-Space Models and Outliers.- 18.1.1 Outliers and Robustness Problems.- 18.1.2 Examples of AO’s and IO’s.- 18.1.3 Problem Setup.- 18.2 Classical Method: Kalman Filter.- 18.2.1 Features of the Classical Kalman Filter.- 18.2.2 Optimality of the Kalman Filter.- 18.3 The rLS filter.- 18.3.1 Derivation.- 18.3.2 Calibration.- 18.3.3 Examples.- 18.3.4 Possible Extensions.- 18.4 The rIC filter.- 18.4.1 Filtering = Regression.- 18.4.2 Robust Regression Estimates.- 18.4.3 Variants: Separate Clipping.- 18.4.4 Criterion for the Choice of b.- 18.4.5 Examples.- 18.4.6 Possible Extensions.- 18.5 Generating Influence Curves.- 18.5.1 Definition of IC.- 18.5.2 General Algorithm.- 18.5.3 Explicite Calculations.- 18.5.4 Integrating along the Directions.- 18.5.5 Auxiliary routines.

O autorze

Wolfgang Härdle is a professor of statistics at the Humboldt-Universität zu Berlin and director of C.A.S.E. the Centre for Applied Statistics and Economics. He teaches quantitative finance and semiparametric statistical methods. His research focuses on dynamic factor models, multivariate statistics in finance and computational statistics. He is an elected ISI member and advisor to the Guanghua School of Management, Peking University and to National Central University, Taiwan.