Mid Term Review Eco 5375 Business and Economic Forecasting Fall 2011 Our upcoming mid term exam is on Monday October 24 2010 at regular class time 5 00 PM in 351 Maguire It will be 1 hour and 20 minutes in length It is a closed notes test and thus you are not to have any notes in any form open during the test Also you must check your phones with the exam proctor You can use any calculators that do not have access to notes and the like We will also provide you with a 4 function calculator if you should need one I recommend that you review your notes your quick quizzes QQ1 QQ7 and your exercises EX1 EX7 Exercises 8 and 9 are due after the mid term and are not a focus of the test The Keys for the quick quizzes and exercises are posted on the course website The major topics for the test are as follows Time Series Decomposition SSP Model Testing for Seasonality Friedman non parametric test Buys Ballot Plots The Box Jenkins Model and Identification methodology The infamous P Q box Forecasting with Box Jenkins Models Augmented Dickey Fuller Unit Root tests The difference in the Prediction Confidence Intervals between BoxJenkins models that assume a stochastic trend and Deterministic Time Series models that assume a deterministic trend Typical seasonal time series transformations used in modeling seasonal time series The Airline Model You should read the following important pdfs and or SAS files that are contained either in the notes section of the course website or in a relevant subdirectory of the course website Review my PowerPoint Introduction to this course Overview ppt Review Some Applications of Forecasting pdf Some Definitions pdf Forecasting Experiments pdf Graphs for out of sample forecasting experiments pdf SAS Lecture pg 1 pdf and SAS Lecture pg 2 pdf Decomposition sas for understanding the basic additive decomposition of time series The results of this program appear in Overview ppt 1 Stable Seasonal Pattern model See Stable Seasonal Pattern Model v2 pdf Buys Ballot plots See the files Buys Ballot Plots pdf and the SAS programs Plano Plot Jacob Williamson sas and Airline Jacob Williamson sas in the Season subdirectory BJ Notation pdf The beginning of the Box Jenkins Model discussion Three Important Concepts pdf Stationarity White Noise and Different Forms of Non stationarity Stats pdf PQ Box Tests for White Noise Residuals and Overfitting In the ACF PACF subdirectory in the course website Study ACF PACF Table pdf and the various graphs of the ACFs and PACFs of the non seasonal Box Jenkins models You should be able to identify various Box Jenkins models simply from these graphs Forecast Profiles pdf the idea of minimum mean squared error forecasts and the role that the conditional mean plays in generating the formulas for such forecasts in the Box Jenkins model You need to understand the outputs from EX4 sas and the EX5 sas program that you were supposed to create Essentially you need to understand the SAS output that generates the statistics needed to fill in the P Q Box Also you should be able to interpret the SAS output produced by Proc ARIMA so that you can write an estimated B J model out in conventional form Augmented Dickey Fuller Unit Root tests You need to be able to interpret output from SAS like in Leadunit sas and DOWlog sas and the EVIEWS output as produced in Lead production wf1 and Dow Jones wf1 Know the different cases of the ADF test See the pdf files ADF Lecture Notes pdf and ADF Notes pdf For some Monte Carlo data sets representing the various ADF cases see Learn Unit Root sas I discussed the logtest SAS macro in the SAS program DOWlog sas Look at the output in that program and be able to use it to decide whether you should work with the data as is or log the data Also you should know of the way to generate level forecasts given the log forecasts and how to create confidence levels of the level variable when given the confidence intervals of the log forecasts See Exercise 7 for more detail You are certainly welcome to go the sub directory Key to Mid terms on the classroom website and look at some of the previous mid term exams in this course You will notice that most of the tests are multiple choice fill in the blank or short answer questions Also note that many of the questions involve the interpretation of SAS output produced by using Proc ARIMA However you should know that some of the topics on these previous mid term exams may not have been covered yet in class as my timing of the presentation of the topics is not always the same from year to year Therefore you should not spend a lot of time studying old questions whose topics have not yet been covered in class Stick with the contents of this Review Sheet and you should be fine 2