本书是预测与时间序列分析课程的教材，书中讲解了预测的重要过程以及可以用于预测的各种统计技术。作者清晰地展示了在营销、金融，人力资源管理，产品调度，过程控制和策略管理中通过预测做出明智决策的重要性。
　　本书适合作为工商管理、理工(包括数学、统计学、计算机科学等)类高年级本科生和研究生的教材，同时可以作为需要进行现实预测的专业人员的参考书。
　　本书的主要特点:
　　清晰、完善地介绍了Box-Jenkins方法精确、易于理解地讨论了传递函数和干涉模型，并介绍了多元时间序列分析给出了基于真实案例的大量习题使用MINITAB和SAS输出给出预测的结果，并有选修的章节详细讲述MINITAB和SAS的用法

Forecasting and Time Series: An Applied Approach, third edition, is designed as a text-book for applied courses in forecasting and time series analysis and as a reference book for practitioners who must make real world forecasts. It is appropriate for advanced (junior-and senior-level) undergraduates and graduate students in business, engineering, and the sciences (including mathematics, statistics, operations research, and computer science). The required mathematical and statistical background for this book is college algebra and basic statistics.
　　This third edition attempts to combine and somewhat expand the best aspects of the first two editions. The first edition began with a short discussion of regression analysis,proceeded to a complete discussion of exponential smoothing and time series decompo-sition techniques, and concluded with a moderately complete presentation of the Box-Jenkins methodology. The second edition began with a more complete treatment of the Box-Jenkins methodology. This methodology was then used to integrate and unify most of the remaining forecasting techniques presented. We have found that some instructors who liked the first edition did not like the organization and Box-Jenkins emphasis of the second. On the other hand, many instructors preferred these aspects of the second edition. The third edition is organized into five parts in a way that should satisfy both those who prefer the first edition and those who prefer the second. Some instructors may wish to teach a course that covers all five sections. However, most courses will not cover the entire text. For instance, some instructors will wish to build their course around the regression/exponential smoothing/decomposition material. Others will wish to build their course around the Box-Jenkins methodology. We now describe the five parts and then explain how they can be structured into different courses.
　　Part I consists of an introduction to forecasting (Chapter 1) and a review of basic statistical concepts (Chapter 2). Part Il discusses forecasting by using regression analysis.
　　This part begins with Chapter 3, which presents simple linear regression. Chapter 4 discusses multiple regression, including an introduction to model building and residual analysis. Chapter 5 presents various advanced topics in regression. These topics are not needed for studying any other chapter in the book, but would be covered by an instruc-tor interested in a more complete discussion of using regression analysis in forecasting.Part III discusses forecasting by using time series regression, decomposition methods,and exponential smoothing. The prerequisite for reading this part is a basic knowl-edge of simple and multiple regression analysis, as provided by Chapters 3 and 4 ofPart II.
　　We suspect that some forecasting courses will have a regression prerequisite, and instructors teaching such courses will begin with Part III. Therefore, we have written this material to stand on its own as much as possible. Part III begins with Chapter 6, which covers time series regression. This includes modeling trends and seasonal effects by using polynomial functions of time, dummy variables, and trigonometric functions.Also covered is an introduction to modeling autocorrelated error terms. Chapter 7 dis-cusses time series decomposition methods. Part III concludes with Chapter 8, which pre-sents exponential smoothing. Included are discussions of simple and double exponential smoothing, Winters' Method, and damped trend methods.
　　Parts IV and V discuss forecasting by using the Box-Jenkins methodology. These parts are written from first principles and can be read without reading Part II or Part III. Therefore, an instructor may begin a course with Part IV, which presents basic techniques of the Box-Jenkins methodology. Chapters 9 and 10 begin this part and discuss nonseasonal Box-Jenkins modeling. Chapter 11 concludes this part and presents an introduction to Box-Jenkins seasonal modeling. It is important to note that, in order to simplify notation, we have delayed use of the backshift operator until Part V. There-fore, the reader can obtain from Part IV a complete knowledge of nonseasonal Box-Jenkins modeling and an introduction to seasonal Box-Jenkins modeling without using this operator. Part V begins with Chapter 12, which presents a more advanced treatment of seasonal Box-Jenkins modeling. Chapter 13 covers the use of the Box-Jenkins meth-odology in time series regression and exponential smoothing. This chapter refers the reader to the necessary prerequisite parts of Chapters 6 and 8 as needed. Part V concludes with Chapter 14, which discusses transfer functions and intervention models.
　　Below we list some possible courses that can be based on this book. All courses are assumed to include the introduction to forecasting provided in Chapter 1 and any needed basic statistical review from Chapter 2.
　　1. A course on forecasting by using regression analysis, time series　regression,　 decomposition methods, and exponential smoothing would consist of Parts II　and III. A more intensive course would also include the basic techniques of the Box-Jenkins methodology, as given by Part IV.
　　2. A course on forecasting by using time series regression, decomposition methods,　　 exponential smoothing, and the basic techniques of the Box-Jenkins methodology would consist of Parts III and IV. A more intensive course would also include the advanced techniques of the Box-Jenkins methodology, as given by Part V.
　　3. A course on forecasting by using the Box-Jenkins methodology would consist of Parts IV and V, with relevant portions of Chapters 6 and 8 included as indicated by the discussions of Chapter 13.
　　4. A course on forecasting by using regression analysis and the Box-Jenkins methodology would consist of Part II, Chapter 6 of Part III, and Part IV. A　more intensive course would also include portions or all of Part V. We have placed a premium in this book on illustrating forecasting by using many real world data sets in the examples and exercises. In addition, we utilize Minitab and SAS outputs to present forecasting results and show in optional sections how to use these packages. A data disk that contains many of the time series in this book is also available.
　　We wish to thank Kathleen Billus, Marcia Cole, Curt Hinrichs, Susan London,Michael Payne, and the other fine people at Duxbury Press for their help in this writing endeavor, as well as Rachel Youngman of Hockett Editorial Service. We would also like to thank the reviewers of this book. We would especially like to thank S. Chakraborti,University of Alabama; Terry Dielman, Texas Christian University; Benito Flores, Texas A & M University; Michael L. Hand, Willamette University; Robert McAuliffe, Babson College; Helmut Schneider, Louisiana State University; Stanley R. Schultz, Cleveland State University; and Mack C. Shelley, II, Iowa State University for their many useful comments and suggestions. Finally, we thank our wives and children for their love and encouragement.
　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　Bruce L. Bowerman
Richard T. O'Connell

CONTENTS
PART　I
INTRODUCTION
CHAPTER　　1
AN　INTRODUCTION TO FORECASTING　　　 2
1.1　Introduction
1.2　Forecasting and Time Series
1.3　Forecasting Methods
1.4　Errors in Forecasting
1.5　Choosing a Forecasting Technique
1.6　An Overview of Quantitative Forecasting Techniques
1.7　Computer Packages: Minitab and SAS
Exercises

CHAPTER　　2
BASIC STATISTICAL CONCEPTS
2.1　Populations
2.2　Probability
2.3　Random Samples and Sample Statistics
2.4　Continuous Probability Distributions
2.5　The Normal Probability Distribution
2.6　The t-Distribution, the F-Distribution, and the Chi-Square Distribution
2.7　Confidence Intervals for a Population Mean
2.8　Hypothesis Testing for a Population Mean
Exercises
PART　　II
FORECASTING　BY　USING　REGRESSION　ANALYSIS

CHAPTER　　3
SIMPLE LINEAR REGRESSION
3.1　The Simple Linear Regression Model
3.2　The Least Squares Point Estimates
3.3　Point Estimates and Point Predictions
3.4　Model Assumptions, the Mean Square Error, and the Standard Error
3.5　Testing the Significance of the Independent Variable
3.6　A Confidence Interval for a Mean Value of the Dependent Variable and a Prediction Interval for an Individual Value of the Dependent Variable
3.7　Simple Coefficients of Determination and Correlation
3.8　An F-Test for the Simple Linear Regression Model
3.9　Using the Computer
Exercises

CHAPTER　4
MULTIPLE　REGRESSION
4.1　The Linear Regression Model
4.2　The Least Squares Point Estimates
4.3　Point Estimates and Point Predictions
4.4　The Regression Assumptions and the Standard Error
4.5　Multiple Coefficients of Determination and Correlation
4.6　An F-Test for the Overall Model
This is an optional section.
4.7　Statistical Inference for Bj and Multicollinearity
4.8　Confidence Intervals and Prediction Intervals
4.9　An Introduction to Model Building
4.10　Residual Analysis
4.11　Using the Computer
Exercises

CHAPTER　　5
TOPICS　IN　REGRESSION　ANALYSIS
5.1　Interaction
5.2　An F-Test for a Portion of a Model
5.3　Using Dummy Variables to Model Qualitative Independent Variables
5.4　Advanced Concepts of Multicollinearity
5.5　Advanced Model Comparison Methods
5.6　Stepwise Regression, Forward Selection, Backward Elimination,and Maximum R2 Improvement
5.7　Outlying and Influential Observations
5.8　Handling Unequal Variances
5.9　Using the Computer
Exercises
III
FORECASTING　BY　USING　TIME SERIES　REGRESSION,
DECOMPOSITION METHODS,
AND EXPONENTIAL SMOOTHING

CHAPTER　　6
TIME SERIES REGRESSION
6.1　Modeling Trend by Using Polynomial Functions
6.2　Detecting Autocorrelation
6.3　Types of Seasonal Variation
This is an optional section
6.4　Modeling Seasonal Variation by Using Dummy Variables
and Trigonometric Functions
6.5　Growth Curve Models
6.6　Handling First-Order Autocorrelation
6.7　Using the Computer
Exercises

CHAPTER　 7
DECOMPOSITION METHODS
7.1　Multiplicative Decomposition
7.2　Additive Decomposition
7.3　Shifting Seasonal Patterns
7.4　The Census II Decomposition Method and SAS PROC X11
7.5　Using the Computer
Exercises

CHAPTER　8
Exponential Smoothing
8.1　Simple Exponential Smoothing
8.2　Adaptive Control Procedures
8.3　Double Exponential Smoothing
8.4　Winters' Method
8.5　Exponential and Damped Trends
8.6　Prediction Intervals
8.7　Concluding Comments
8.8　Using the Computer
Exercises
PART　IV
FORECASTING BY USING BASIC TECHNIQUES
OF THE　BOX-JENKINS　METHODOLOGY
This is an optional section.
CONTENTS

CHAPTER　9
NONSEASONAL BOX-JENKINS MODELS
AND THEIR TENTATIVE IDENTIFICATION
9.1　Stationary and Nonstationary Time Series
9.2　The Sample Autocorrelation and Partial
Autocorrelation Functions: The SAC and SPAC
9.3　An Introduction to Nonseasonal Modeling and Forecasting
9.4　Tentative Identification of Nonseasonal Box-Jenkins Models
9.5　Using the Computer
Exercises

CHAPTER　10
ESTIMATION, DIAGNOSTIC CHECKING, AND FORECASTING
FOR NONSEASONAL BOX-JENKINS MODELS
10.1　Estimation
10.2　Diagnostic Checking
10.3　Forecasting
10.4　A Case Study
10.5　Using the Computer
Exercises

CHAPTER　 11
AN INTRODUCTION TO
BOX-JENKINS SEASONAL MODELING
11.1　Transforming a Seasonal Time Series into a Stationary Time Series
11.2　Two Examples of Seasonal Modeling and Forecasting
11.3　Using the Computer
Exercises

PART　　V
FORECASTING　BY USING　ADVANCED TECHNIQUES OF
THE　BOX-JENKINS　METHODOLOGY
This is an optional
　　　　　
CHAPTER
GENERAL BOX-JENKINS SEASONAL MODELING
12.1　The General Seasonal Model and Guidelines for Tentative Identification
12.2　Improving an Inadequate Seasonal Model
12.3　Using the Computer
Exercises

CHAPTER
USING THE BOX-JENKINS METHODOLOGY TO
IMPROVE TIME SERIES REGRESSION MODELS
AND TO IMPLEMENT EXPONENTIAL SMOOTHING
13.1　Box-Jenkins Error Term Models in Time Series Regression
13.2　Seasonal Intervention Models
13.3　Box-Jenkins Implementation of Exponential Smoothing
13.4　Using the Computer
Exercises

CHAPTER　 14
TRANSFER FUNCTIONS AND INTERVENTION MODELS
14.1　A Three-Step Procedure for Building a Transfer Function Model
14.2　Intervention Models
14.3　Using the Computer
Exercises
APPENDIX　　A
STATISTICAL TABLES
APPENDIX　　B
REFERENCES
Index
This is an optional section