課程資訊
課程名稱
 應用貝氏統計分析
Applied Bayesian Statistical Analysis 
開課學期
 107-1 
授課對象
 管理學院  國際企業學研究所  
授課教師
 任立中 
課號
 IB7091 
課程識別碼
 724 M4210 
班次
  
學分
 2.0 
全/半年
 半年 
必/選修
 選修 
上課時間
 星期一A,B(18:25~20:10) 
上課地點
 管一405 
備註
 限碩士班以上
總人數上限：20人
外系人數限制：5人 
Ceiba 課程網頁
 http://ceiba.ntu.edu.tw/1071IB7091_ 
課程簡介影片
 
核心能力關聯
 核心能力與課程規劃關聯圖
課程大綱
為確保您我的權利,請尊重智慧財產權及不得非法影印
課程概述

The purpose of this course is to provide the advance topics in statistical analysis, Bayesian statistics, and time-series forecasting through examples and with an emphasis on applications in managerial decision, real data, and the use of statistical software and spreadsheets. Because of the practical value of these statistical techniques, the course takes great care to explain the proper application of the methods and the importance of correctly interpreting computer output. Bayesian statistical analysis is one of the most important tools for management. Based on what we have learned in the introductory course of Statistics, some advance issues, which are closer to the real world problems such as distribution theory, multicollinearity, lagged variables, heteroscedasticity, autocorrelation, and limited dependent variables, will be discussed in details.  

課程目標
 Bayesian statistics has developed a great deal over the past 20-year period in the fields of marketing science especially. We will introduce the fundamental concepts of Bayesian statistics and its applications to the modern big data marketing or customer relationship marketing analysis. Finally, the last part of this course will cover the statistical analysis for time series data that is an essential technique for financial decision-making. 
課程要求
 COURSE PREREQUISITES:

1. Management
2. Statistics
 
預期每週課後學習時數
  
Office Hours
 每週一 14:00~16:00 
指定閱讀
 TEXT BOOKS:

1. Lynch, M. Scott (2007), Introduction to Applied Bayesian Statistics and Estimation for Social Scientists, Springer Science+Business Media, LLC, New York, NY. (e-books in NTU library)
 
參考書目
 REFERENCE BOOKS:

2. Rossi, Peter E., Greg M. Allenby, and Rob McCulloch (2005), Bayesian Statistics and Marketing, John Wiley and Sons, New York, NY.
 
評量方式
(僅供參考)
  
No.
項目
百分比
說明
1. 
 Homework Assignments  
25% 
  
2. 
 Midterm Exam 
25% 
 Midterm Exam — (November 5) 
3. 
 Final Exam 
25% 
 Final Exam — (January 7) 
4. 
 Class Participation 
25% 
  
 
課程進度
週次
日期
單元主題
第1週
 9/10  Introduction to Course Philosophy, Structure, and Policy
Lecture: 1. Outline
2. A note on programming
3. Symbols used throughout the course
Discussion: Free discussion
Assignments: 1. Preview Chapter 1 Introduction 
第2週
 9/17  Foundations: Review of Basic Statistical Concepts
Lecture: 1. Rules of probability
2. Probability distribution in general
3. Classical statistics and Maximum Likelihood Estimation
Discussion: Free discussion
Assignments: 1. Gelman, Andrew (2008), “Teaching Bayes to Graduate Students in Political Science, Sociology, Public Health, Education, Economics, …,” The American Statistician, 62:3, 202-205.
2. Preview Chapter 2 Probability Theory and Classical Statistics 
第3週
 9/24  The Mid-Autumn Festival Holiday 
第4週
 10/01  Foundations (I): Discrete Distribution Theory
Lecture: 1. Rules of probability
2. Probability distribution in general
3. Classical statistics and Maximum Likelihood Estimation The Binomial Distribution
4. The Multinomial Distribution
5. The Poisson Distribution
Discussion: Gelman (2008)
Assignments: 1. Allenby, Greg M. and Peter E. Rossi (2008), “Teaching Bayesian Statistics to Marketing and Business Students,” The American Statistician, 62, 3, 195-198.
2. Preview Chapter 2 Probability Theory and Classical Statistics
 
第5週
 10/08  Foundations (III): Continuous Distribution Theory
Lecture: 1. The Beta/Dirichlet Distribution
2. The Normal Distribution
3. The Multivariate Normal Distribution
4. The Generalized Gamma Distribution
Discussion: Allenby and Rossi (2008)
Assignments: 1. 任立中，林婷鈴，陳靜怡，李吉仁，2006年，高科技產業產品價值創造與行銷價值專屬化之最適資源配置，中山管理評論，第十四卷第一期，第11至42頁。
2. Preview Chapter 3 Basics of Bayesian Statistics
 
第6週
 10/15  Bayesian Inference (I): Sample, Prior, and Posterior
Lecture: 1. Bayes’ Theorem for point probabilities
2. Bayes’ Theorem applied to probability distributions
3. Bayes’ Theorem with distributions: A voting example
4. Criticism against Bayesian statistics
Discussion: 任立中，林婷鈴，陳靜怡，李吉仁 (2006)
Assignments: 1. Jen, Lichung and Han-Kuang Tien (2013), “Ascertaining the Dynamic Competition in Channel Relationship Management,” International Journal of Marketing Studies, Vol. 5, Issue 3, 36-47.
2. Preview Chapter 3 Basics of Bayesian Statistics
 
第7週
 10/22  Bayesian Inference (II): Beta–Binomial Model
Lecture: 1. Beta–Binomial Model and Its Applications
2. Dirichlet–Multinomial Model and Its Application
Discussion: Jen and Tien (2013)
Assignments: 1. Jen, Lichung, Chien-Heng Chou, and Greg M. Allenby (2003), “A Bayesian Approach to Modeling Purchase Frequency,” Marketing Letters, Vol. 14, No. 1, 5-20.
2. Preview Chapter 3 Basics of Bayesian Statistics
 
第8週
 10/29  Bayesian Inference (III): Poisson–Gamma Model
Lecture: 1. Poisson—Gamma Model and Its Application
Discussion: Jen, Chou, and Allenby (2003)
Assignments: 1. Allenby, Jen, and Leone (1996), “Economic Trends and Being Trendy: The Influence of Consumer Confidence on Retail Fashion Sales,” Journal of Business & Economic Statistics, 14(1), 103-111.
2. Preview Chapter 3 Basics of Bayesian Statistics
 
第9週
 11/05  Midterm Exam – Chapter 1~6 (25%) 
第10週
 11/12  Bayesian Inference (IV): Normal-Normal-Inverted Gamma Model
Lecture: 1. Normal—Normal Model and Its Applications
2. Normal—Inverted Gamma Model and Its Application
Discussion: Allenby, Jen, and Leone (1996)
Assignments: 1. Preview Chapter 4 Gibbs Sampling 
第11週
 11/19  Modern Model Estimation Part 1: Gibbs Sampling
Lecture: 1. What Bayesians want and why
2. The logic of sampling from posterior densities
3. Two basic sampling methods
4. Introduction to MCMC sampling
Discussion: Free discussion
Assignments: 1. Edwards, Yancy D. and Greg M. Allenby (2003) "Multivariate Analysis of Multiple Response Data," Journal of Marketing Research, 40, 321-334.
2. Preview Chapter 10 Introduction to Multivariate Regression Models
 
第12週
 11/26  Multivariate Regression
Lecture: 1. Model development
2. Implementing the algorithm Hierarchical models in general
3. Multivariate probit models
Discussion: Edwards and Allenby (2003)
Assignments: 1. Allenby, Greg M., Neeraj Arora and James L. Ginter (1995) "Incorporating Prior Knowledge into the Analysis of Conjoint Studies," Journal of Marketing Research, 32, 152-162.
2. Preview Chapter 9 Introduction to Hierarchical Models 
第13週
 12/03  HB Regression: Interaction Model
Lecture: 1. Hierarchical models in general
2. Hierarchical linear regression models
3. Random effects: The random intercept model
4. Random effects: The random coefficient model
5. Growth models
6. A note on fixed versus random effects models and other terminology
Discussion: Allenby, Arora and Ginter (1995)
Assignments: 1. Siddhartha Chib and Edward Greenberg (1995) "Understanding the Metropolis-Hastings Algorithm," The American Statistician, 49, 4, 327-335.
2. Preview Chapter 5 Metroplis–Hastings Sampling
 
第14週
 12/10  Modern Model Estimation Part 2: Metroplis–Hastings Sampling
Lecture: 1. Relationship between Gibbs and MH sampling
2. Example: MH sampling when conditional densities are difficult to derive
3. Example: MH sampling for a conditional density with an unknown form
4. Extending the bivariate normal example: The full multiparameter model
Discussion: Chib and Greenberg (1995)
Assignments: 1. Allenby, Greg M. and Peter E. Rossi (2008), “Teaching Bayesian Statistics to Marketing and Business Students,” The American Statistician, 62, 3, 195-198.
2. Preview Chapter 8 Generalized Linear Models  
第15週
 12/17  Revealed Preference Models
Lecture: 1. The Dichotomous Probit Model
2. The Ordinal Probit Model
Discussion: Allenby, Jen, and Leone (1996)
Assignments: 1. Jen, Lichung, Demetrios Vakratsas , and Wei-Lin Wang (2012), “Accounting for Deviations from Routine Timing Behavior: An Individual-Level Approach,” Working paper, NTU.
2. Liu, Hsiu-Wen (2007), “Hierarchical Bayes Conjoint Analysis with Multivariate Mixture of Normal Heterogeneity: Implications for Market Segmentation,” Doctoral dissertation, NTU. 
第16週
 12/24  Mixture Model
Lecture: 1. The Mixture Normal Model and Its Application
2. The Mixture Gamma Model and its Application
Discussion: Jen, Vakratsas, and Wang (2012), Liu (2007)
Assignments: 1. Preview Chapter 6 Evaluating Markov Chain Monte Carlo Algorithms and Model Fit
 
第17週
 12/31  Evaluating Markov Chain Monte Carlo Algorithms and Model Fit
Lecture: 1. Why evaluate MCMC algorithm performance?
2. Some common problems and solutions
3. Recognizing poor performance
4. Evaluating model fit
5. Formal comparison and combining models 
第18週
 1/7  Final Exam (25%)