Fundamentals of Forecasting Using Excel

Dr. Kenneth D. Lawrence

Dr. Ronald K. Klimberg
Dr. Sheila M. Lawrence

Industrial Press, Inc.

New York, New York
 Contents

About the Authors xi

Chapter 1 INTRODUCTION TO FORECASTING

1.1 Introduction 1
1.2 Forecasting 2
1.2.1 Importance 2
1.2.2 Four Forecasting Factors to Consider 2
1.2.3 Forecasting Method Decision Tree 3
1.2.4 Forecasting Techniques 4
1.3 Excel 5
1.3.1 Installing the Data Analysis ToolPak 6
References 8

Chapter 2 SUMMARIZING AND DISPLAYING DATA I N EXCEL . ,9

2.1 Introduction 9
2.2 Summarizing Continuous Data (Descriptive2.xls and Boxplot.xls) 10
2.2.1 Descriptive Statistics (Descriptive2.xls\Worksheet: Descriptive) 10
2.3 Graphing Continuous Data (Descriptive, Worksheets: Histograml, Histogram2, and Time Series Plot) 12
2.3.1 Box plot (Descriptive2: Worksheet: Descriptive; Boxplot.xls: Worksheet: Simple) 12'
2.3.2 Histogram 13
2.3.3 Time Series Plot (Worksheet: Time Series Plot) 16
2.4 Summarizing Qualitative and Discrete Data; Graphs and Tables with Qualitative and Discrete Data
(Descriptive2.xls) 17
VI Contents

2.4.1 Descriptive Statistics (Descriptive2; Worksheet: Countif) 17

2.4.2 Graphing Qualitative and Discrete Data 17
2.4.3 PivotTables with a Row or Column being a Continuous Variable 21
2.5 A General Hypothesis Testing Framework 21
2.5.1 One-sample hypothesis testing of the mean (Continuous data) 23
2.6 Excel Nuggets (Datatable2.xls) 26
2.6.1 Scroll bar and One and Two-way Data Tables (Worksheets: Mortl-4) 26
2.6.2 Goal Seek (Worksheet: Goal) 29
References 31

Chapter 3 BASIC T I M E SERIES ANALYSIS 33

3.1 Introduction 33
3.2 Components of a Time Series (Files: Datats.xls and Randomness.xls) 34
3.2.1 Trend Component (Worksheet: Trend) 34
3.2.2 Linear Trend Analysis 34
3.2.3 Test for Significant Trend 37
3.2.4 Seasonal Component (Worksheet: Seasonal) 38
3.2.5 Cyclical Component 40
3.2.6 Irregular Component (Randomness.xls) 41
3.3 Decomposition (Datats.xls) 41
3.3.1 Multiplicative Decomposition (without cycle component) (Worksheet: Decompl) 42
3.3.2 Multiplicative Decomposition (with cycle component) 45
3.4 Moving Averages (File: Datats.xls, Worksheet: MA) 48
3.5 Exponential Smoothing (Worksheet: EA(.l)) 50
3.5.1 Optimal Smoothing Constant 54
3.5.2 Summary 55
References 55

Chapter 4 FORECASTING PERFORMANCE MEASUREMENTS, TRACKING SIGNALS, AND RANDOMNESS TESTS... 57

4.1 Introduction 57
4.2 Forecast Performance Measures 58
4.3 Tracking Signals 61
4.4 Randomness 61
References 68
 Fundamentals of Forecasting Using Excel Vll

Chapter 5 ADVANCED TIME SERIES FORECASTING TECHNIQUES 69

5.1 Introduction 69
5.2 Advanced Time Series Techniques 70
5.2.1 Double Moving Average 70
5.2.2 Double Exponential Smoothing 73
5.2.3 Brown's Double Exponential Smoothing 73
5.2.4 Holt's Double Exponential Smoothing 74
5.2.5 Exponential Smoothing with a Seasonal Component and with/without a Trend 76
5.3 Autoregressive Integrated Moving Average (ARIMA) Models 79
5.3.1 Autocorrelation 79
5.3.2 Differencing 83 -
5.3.3 Autoregressive and moving average models 84
5.3.4 ARIMA models 85
5.4 Index Numbers 87
5.4.1 Deflating data 88
5.4.2 Aggregate Index numbers 90
References 91

Chapter 6 SIMPLE LINEAR REGRESSION ANALYSIS IN FORECASTING 93

6.1 Introduction 93
6.2 The Simple Linear Regression Model 94
6.3 Assumptions Underlying the Simple Linear Regression Model 94
6.4 Standard Error of the Estimate 95
6.5 Predicting Y 95
6.6 Correlation 95
6.7 Coefficient of Determination 96
6.8 Residuals 96
6.9 Statistical Inference on the Slope of the Regression Model 96
6.10 Testing for First-Order Autocorrelation 98
6.10.1 The Durbin-Watson Statistic 98
6.11 An Example of Simple Linear Regression 99
6.12 Hypothesis Testing 102
6.13 Example Problems in Excel 2007 for Simple Linear Regression Forecasting 103
Vlll Contents

6.13.1 Regression AnalysisProblem 1 in Excel 2007 103

6.13.2 Solution of Problem 1 104
6.13.3 Regression AnalysisProblem 2 in Excel 2007 106
6.13;4 Solution of Problem 2 106
6.13.5 Regression Charts 108
6.13.6 Regression AnalysisProblem 3 in Excel 2007 110
6.13.7 Solution of Problem 3 111
6.13.8 Polynomial Model 114
6.13.9 Logarithmic Model 118
6.13.10 Power Model 119
6.13.11 Exponential Model 121
6.13.12 Comparison of the Models 122
6.14 Example Problems in Excel 2007 for Simple Linear Regression Forecasting 122
6.14.1 AutocorrelationA Common Problem with Time Series Data 122
6.14.2 Heteroscedasticity: Unequal Error VarianceIdentification and Remedy 124
6.14.3 Solution of Heteroscedasticity Example 125
6.15 Regression AnalysisSummary Output and Residuals 126
References 131

Chapter 7 MULTIPLE LINEAR REGRESSION ANALYSIS IN FORECASTING 133

7.1 Introduction 133
7.2 The Multiple Regression Model 134
7.3 The Multiple Regression Model in Matrix Terms 134
7.4 Analysis of Variance 135
7.5 F Test for the Multiple Regression Model 136
7.6 Coefficient of Multiple Determination 136
7.7 Inference about the Regression Parameters 136
7.8 Estimation and Prediction 137
7.9 Diagnostics and Their Use in Multiple Regression Models 137
7.9.1 Scatter Plots 137
7.9.2 Residual Plots 137
7.10 An Example Problem 137
7.11 Another Example Problem 138
7.12 Qualitative Variable 139
 Fundamentals of Forecasting Using Excel IX

7.13 Polynomial Regression 139

7.14 Nonlinear Multiple Regression Models 140
7.14.1 Multiplicative Model 140
7.14.2 Exponential Model 140
7.14.3 Reciprocal Model 141
7.15 Variance Stabilizing Transformation 141
7.15.1 Square Root Transformation 141
7.15.2 Logarithmic Transformation 141
7.15.3 Reciprocal Transformation 141
7.16 Multiple Regression Models with Indicator Variables 141
7.17 Multicollinearity 1 4 2 -
7.17.1 Correcting the Problems of Multicollinearity 142
7.18 Multiple Regression Examples 143
7.18.1 Multiple RegressionExample 1 143
7.18.2 For Further Insight 150
7.18.3 Multiple Regression Example 2: Using Categorical Variables 151
References 155

Chapter 8 MARKOVIAN FORECASTING MODELS 157

8.1 Introduction 157
8.2 A Markovian Model 158
8.3 The First-Order Markovian Model 158
8.4 Computation of the Steady State 159
8.4.1 An Example Problem for Steady State Calculations 159
8.4.2 Problem 1Brand Calculations Steady State 160
8.5 Absorbing States in the Markov Process 162
8.5.1 Problem 2Absorbing States Calculations for Credit 163
8.5.2 Problem 3Inventory Questions 165
References 166

Chapter 9 DIFFUSION MODELING AND FORECASTING 167

9.1 Introduction 167
9.2 Diffusion Model 169
9.3 Another Model for Estimators: The Product Life Cycle Using Excel Computation 170
X Contents

9.4 Excel ComputationsForecasting 170

9.5 A Markovian Customer Demand Level Model and the Diffusion Model 172
References 173

Chapter 10 MISCELLANEOUS FORECASTING METHODS 175

10.1 Gravity Modeling 175
10.1.1 Introduction 175
10.1.2 Journey-to-Work Model and the Need for Gravity Model Constraints 176
10.1.3 Estimation of the Model Parameters 178
10.2 Input-Output Modeling 180
10.2.1 Introduction 180
10.2.2 Regional Input-Output Models 182
10.2.3 An Example Problem 183
10.3 Combination of Forecasts 185
10.3.1 Introduction 185
10.3.2 Methods of Combining 186
10.4 Combined Forecasting Method 187
References for the Gravity Model 190
References for Input-Output Modeling 190
References for Combination of Forecasts 191

Index 193