| |
|
|
|
Syllabus |
|
|||||||||||
|
||||||||||||
Four items are needed to begin this course:1) Textbook: J. Holton Wilson and Barry Keating. Business Forecasting, Sixth Edition (McGraw-Hill/Irwin, 2009) ISBN 978-0073373645
For users of iPads, tablets, laptops, or any device able to display an Internet Browser, you may access the textbook digitally from VitalSource. Click the icon below and search on "Keating."
2) ForecastX For Excel (statistical software downloaded with instructions emailed to you and also available in the Mendoza labs); this software is compatible with Windows 7, Windows 10 and Office 2016 (as well as with Office 2013). This software will be made available for download to registered students. This software is only available for Windows.
3) XLMiner (not available on any college machines or servers). This software will be available for download to registered students. This software is only available for Windows.
4) Prevedere (accessed through a browser). The Prevedere application will be available directly to registered students.
Other software we will use is available on College servers: IBM/SPSS Modeler and SAS Enterprise Miner; these commercial packages are not available for student download.
"I believe that forecasting or demand management may have the potential to add more value to a business than any single activity within the supply chain. I say this because if you can get the forecast right, you have the potential to get everything else in the supply chain right. But if you can't get the forecast right, then everything else you do essentially will be reactive, as opposed to proactive planning."
Al Enns, Director of Supply Chain Strategies, Motts North America, Stamford, Connecticut
We might think of business prediction as having developed in three stages. One way of looking at this development is illustrated in the figure below. Almost anything you think about has gone through some evolution. In lighting, we went from primitive torches to kerosene to the first light bulbs developed by Edison in the late 1870s. Incandescent light bulbs like the one at the left in the figuure became a worldwide standard for many years. Such light bulbs had a relatively short life, were hot, and were not energy efficient. In the late 20th century, new, more energy-efficient, longer lasting, but still hot compact flourescent bulbs became common. Then, in the early 21st century, LED bulbs that were cool, energy efficient, and long-lasting became the new norm.
Forecasting has gone through an evolution as well. The oldest of the predictive models, time series models, are covered at the beginning of the course (Stage I). The second stage of the course moves into causal demand planning models (Stage II). But the emergence of "Big Data" has brought with it an increased need for additional analytical tools; these analytics tools will comprise the last third of the course. Today every organization is confronted with a deluge of data. Organizations have more data than can easily be converted into actionable information using traditional statistical methods. This is "Stage III" in the evolution of forecasting and predictive analytics. The single goal of this course is to enable you to turn "data" into "information." Data is simply numbers, text, audio, and graphics; data is just facts and figures. Information, however, is data that is processed, organized, structured or presented in context so as to make it useful for informing decisions.
There is a certification process available to forecasters much like the Certified Financial Analyst designation or the Certified Public Accountant designation. The Certified Professional Forecaster designation is available through the Institute of Business Forecasting.
Course Learning Objectives:
If you stay engaged in the course and complete the suggested readings and assignments:
Attendance:
Regular attendance is essential to the successful completion of this course. Attendance will regularly be taken and you are responsible for material covered in class whether or not you have attended class. Missing more than two class sessions (for any reason) will result in an automatic reduction in course grade. Unsatisfactory attendance may result in a failing grade. You should plan on spending at least two hours of independent study for each hour of class attendance.
Grading:
A course grade will be assigned on the basis of student performance on two examinations, a final examination, assignments, and textbook problems. The assignments and textbook problems will be presented by groups in class.
Class presentation (by group) of assigned exercises (problems): twenty percent of the course grade
Midterm Exam : forty percent of the course grade
Final (comprehensive) Exam : forty percent of the course grade
The Gradebook is available in Sakai.
Assignments and Problems:
On the attached "assignment sheet" you will find a class-by-class list of topics to be covered and your reading assignment. Reading assignments in the textbook are to be completed before the class day under which they are listed in the assignment sheet. Problem assignments are to be completed on the date following the listing and the solutions may be presented by the selected group spokesperson to the class. It is the group's presentation of the assigned problems in class that counts as participation; problems are rarely handed in as hardcopy. Not all the exercises and problems listed in the syllabus will be due; the instructor will assign the particular problems and exercises due after the completion of each class meeting.
Assignments (essentially longer problems, directed exercises, or reviews of articles presented in class) will be assigned for most of the topics covered and will be presented by students in class. The class presentation of assignments and textbook problems (using the computer) by each group is an important and integral part of the course. Assignments are drawn from the "Assignment Sheet" below.
Honor Code:
The University of Notre Dame Graduate Academic Code of Honor is observed in this class. Violation of the Code of Honor consists of misrepresenting, in any way, anyone else’s work as your own, any verbal or written misrepresentations to the instructor, any use of unauthorized external materials during quizzes and/or tests, or any collaborative effort on the examinations. All members of the class have an equal and shared responsibility to enforce the code of ethics among their peers.
Midterm Examination:
The examination will be a full-period examination of essentially a problem-solving nature; problems will be similar to those in the textbook. Because of the technical nature of the examination, students are allowed to use calculators.These examinations, however, are to be completed without reference to the textbook, class notes or any other materials. The University of Notre Dame Graduate Academic Code of Honor is in effect for these examinations. No RF (i.e., radio frequency) devices are allowed to be used during an examination; possession of such a device during an examination is an automatic Honor Code violation (this includes cell phones). Cell phones may be placed on a table at the head of the classroom before the exam begins. The test may also include a practicum using the forecasting and predictive analytics modeling software assigned for class use. This midterm examination will be returned to you after grading (as have midterms in past semesters).
Final Comprehensive Examination:
A comprehensive final examination will be administered during the "final examination period" of the university at the Registrar's selected time and date (see below). The same procedures which apply to the midterm examinations also apply to this examination. The University of Notre Dame Graduate Academic Code of Honor is again in effect for this examination.
The Project:
Note: This semester graduate students will not have a project!
Missing Assignments:
Assignments or problems not ready for presentation on the due date (that is the instructor assigned due date) will receive a grade of zero. It is your group's presentation of the assignments and the problems that are graded.
Class # Date Topic Assignment
This course meets on Mondays and Wednesdays
1 10/23 (Monday) Introduction to Business Forecasting - Chapter 1
- Overview of the ForecastXTM computing package, the XLMinerTM analytics package, SAS Enterprise Miner, and IBM/SPSS Modeler
- The Syllabus - What we will be doing, and what we will not be doing
Pre Class Assignment - Please complete before first class meeting Chapter1, Exercise 8
Use ForecastX
Chapter 2, Exercise 9
Use ForecastX
Chapter 3, Exercise 6
Use ForecastX
2 10/25 (Wednesday) The Forecast Process, Data Considerations, and Model Selection -- Chapter 2 and Moving Averages and Exponential Smoothing -- Chapter 3
New Product Forecasting and Diffusion Models
Condiment I Problem (do not include "events" in the analysis) Condiment II Problem (include "events" in the analysis) Disinfectant I Problem (do not include "events" in the analysis) Disinfectant II Problem (include "events" in the analysis) CD Adoptions - Use Intermittent Method (Use both Croston's Method and the Slow Moving Method) Cellular Telephone Adoption - Try a Bass Model (0.008, 0.421, 100) Color Television Adoption - Try a Gompertz Model HDTV Adoption - Try a Gompertz Model Microelectronic Density Forecast - Use a Gompertz Model World Nuclear Generating Capacity - Use a Logistics Model
3 10/30 (Monday) -- Introduction to Forecasting with Regression Methods --Chapter 4
problem c4p4 problem c4p5 problems c4p6 problem c4p7 problem c4p8 problem c4p9 problem c4p10 problem c4p11 problem c4p12 problem c4p13 problem c4p14
Create a "growth model" with original data.
4 11/1 (Wednesday) -- Forecasting with Multiple Regression and the use of "Outside data" and Prevedere -- Chapter 5
Note: Read Chapter 5 in the current textbook; however, the instructor will also pass out an addition to Chapter 5 covering the Prevedere material.
problem c5p5 problem c5p6 problem c5p7
problem c5p8 problem c5p9 problem c5p10 (use the "Economagic" site to collect data) problem c5p11 problem c5p12 problem c5p13 Create a causal multiple regression with original data.
5 11/6 (Monday) -- Time-Series Decomposition --Chapter 6 and Box-Jenkins (ARIMA) Type Forecasting Models -- Chapter 7
problem c6p5 problem c6p6 problem c6p7 problem c6p8 problem c6p11 problem c6p6 problem c6p9 problem c6p12
problem c7p5 problem c7p6 problem c7p8 problem c7p9
6 11/8 (Wednesday) -- Combining Forecast Results (Ensemble Models) - Chapter 8
problem c8p3 problem c8p4 problem c8p5 problem c8p6
7 11/13 (Monday) -- Introduction to Predictive Analytics and Data Mining with XLMinerTM -- Chapter 9
Note: Do read Chapter 9 in the current textbook; however, the instructor will pass out hardcopy of four new chapters covering the Predictive Analytics section of the course.
8 11/15 (Wednesday) Midterm Examination
9 11/20 (Monday) -- Classification and k-Nearest Neighbor Models -- Chapter 9
Note: Do read Chapter 9 in the current textbook; however, the instructor will pass out hardcopy of four new chapters covering the Predictive Analytics section of the course.
| Missing Values Exercise | Partition Exercise |
"Confusion Matrix" Explanation "Lift Chart" Explanation K-Nearest Neighbor Exercise #1
K-Nearest Neighbor Exercise #2
Use k-nearest neighbor analysis on the following data sets:
ridingmowers universalbank accident
Thanksgiving Holiday is Wednesday November 22 - Sunday November 26
10 11/27 (Monday) -- CART, Naive Bayes Models, and Logit-- Chapter 9
Note: Do read Chapter 9 in the current textbook; however, the instructor will pass out hardcopy of four new chapters covering the Predictive Analytics section of the course.
Use CART on the following data:
Use Naive Bayes on the following data:
Naive Bayes Titanic Exercise Use Logistics Regression analysis on the following data sets:
Rainy Days Logistics Regression Logistics Regression Titanic 2 Exercise (uses a different data set than the exercise above - passengers only, no crew) Mail Order Customers Logistic Regression Universal Bank for SAS Enterprise Miner Analysis
11 11/29 (Wednesday) -- Introduction to Text Analytics ("Bag of Words" algorithm as used in XLMiner)
12 12/4 (Monday) -- Text Analytics ("Natural Language Processing" algorithm as used in IBM/SPSS Modeler) - Capturing Social Media Sentiment as It Happens
13 12/6 Last Class Day - (Wednesday) -- -- Text Analytics Concluded
Examination for Business Forecasting (FIN 70230):
Monday December 11, 2017
Location: L068
Time: 8:00 - 9:50 am
