Multivariate Methods and Forecasting with IBM® SPSS® Statistics (eBook)

Abdulkader Aljandali, Ph.D., is Senior Lecturer at Regent’s University London. He currently leads the Business Forecasting and the Quantitative Finance module at Regent’s in addition to acting as a Visiting Professor for various universities across the UK, Germany and Morocco. Dr Aljandali is an established member of the Higher Education Academy (HEA) and an active member of the British Accounting and Finance Association (BAFA).

Preface 6
Introduction 8
Contents 10
List of Figures 13
List of Tables 17
Part I: Forecasting Models 18
Chapter 1: Multivariate Regression 19
1.1 The Assumptions Underlying Regression 20
1.1.1 Multicollinearity 20
1.1.2 Homoscedasticity of the Residuals 21
1.1.3 Normality of the Residuals 24
1.1.4 Independence of the Residuals 24
1.2 Selecting the Regression Equation 27
1.3 Multivariate Regression in IBM SPSS Statistics 28
1.4 The Cochrane-Orcutt Procedure for Tackling Autocorrelation 35
Chapter 2: Other Useful Topics in Regression 42
2.1 Binary Logistic Regression 43
2.1.1 The Linear Probability Model (LPM) 43
2.1.2 The Logit Model 46
2.1.3 Applying the Logit Model 47
2.1.4 The Logistic Model in IBM SPSS Statistics 48
2.1.5 A Financial Application of the Logistic Model 54
2.2 Multinomial Logistic Regression 55
2.3 Dummy Regression 55
2.4 Functional Forms of Regression Models 62
2.4.1 The Power Model 64
2.4.2 The Reciprocal Model 67
2.4.3 The Linear Trend Model 70
Chapter 3: The Box-Jenkins Methodology 73
3.1 The Property of Stationarity 73
3.1.1 Trend Differencing 74
3.1.2 Seasonal Differencing 76
3.1.3 Homoscedasticity of the Data 77
3.1.4 Producing a Stationary Time Series in IBM SPSS Statistics 77
3.2 The ARIMA Model 80
3.3 Autocorrelation 81
3.3.1 ACF 81
3.3.2 PACF 84
3.3.3 Patterns of the ACF and PACF 85
3.3.4 Applying an ARIMA Model 85
3.4 ARIMA Models in IBM SPSS Statistics 88
Chapter 4: Exponential Smoothing and Naïve Models 94
4.1 Exponential Smoothing Models 94
4.2 The Naïve Models 101
Part II: Multivariate Methods 107
Chapter 5: Factor Analysis 108
5.1 The Correlation Matrix 109
5.2 The Terminology and Logic of Factor Analysis 109
5.3 Rotation and the Naming of Factors 113
5.4 Factor Scores in IBM SPSS Statistics 116
Chapter 6: Discriminant Analysis 118
6.1 The Methodology of Discriminant Analysis 118
6.2 Discriminant Analysis in IBM SPSS Statistics 119
6.3 Results of Applying the IBM SPSS Discriminant Procedure 121
Chapter 7: Multidimension Scaling (MDS) 128
7.1 Types of MDS Model and Rationale of MDS 130
7.2 Methods for Obtaining Proximities 131
7.3 The Basics of MDS in IBM SPSS Statistics: Flying Mileages 132
7.4 An Example of Nonmetric MDS in IBM SPSS Statistics: Perceptions of Car Models 137
7.5 Methods of Computing Proximities 138
7.6 Weighted Multidimensional Scaling in IBM SPSS, INDSCAL 141
Chapter 8: Hierchical Log-linear Analysis 145
8.1 The Logic and Terminology of Log-linear Analysis 145
8.2 IBM SPSS Statistics Commands for the Saturated Model 148
8.3 The Independence Model 152
8.4 Hierarchical Models 154
8.5 Backward Elimination 158
Part III: Research Methods 160
Chapter 9: Testing for Dependence 161
9.1 Introduction 161
9.2 Chi-Square in IBM SPSS Statistics 163
Chapter 10: Testing for Differences Between Groups 167
10.1 Introduction 167
10.2 Testing for Population Normality and Equal Variances 168
10.3 The One-Way Analysis of Variance (ANOVA) 170
10.4 The Kruskal-Wallis Test 172
Chapter 11: Current and Constant Prices 175
11.1 HICP and RPI 175
11.2 Current and Constant Prices 176
References 181
Index 182