Please put all of your work into a single .doc, .pdf or .ppt document. Presentation is not a part of the assessment, but please make it easy for me to find all the pieces of your work.
Part I : Find a paper that uses an ARIMA or GARCH model for any forecasting application. (Searching JSTOR or Google Scholar for ARIMA model is a quick way to get hits, but any source – academic or industry – is fine).
Part 1 Submission: Create 3-5 slides presenting an overview of the paper. Please make sure to cover the points listed below.
a)Cite the article’s title, author(s) and publication title and date.
b)Describe the class(es) of models that the authors built.
c)Describe the methodology that was used to identify the order of the model.
d)Describe the paper’s major findings.
e)Describe 2 questions that the paper raises for you, i.e., ideas for further research, disagreements you have with the authors conclusion, a different approach to the paper’s subject, a different subject of your interest for which the paper’s methodology could be adapted.
Part II : Choose one of the 3 options below and build your own forecasting model. Your model can be a replication or extension of the paper that you found for Part I, but it does not have to be related.
Option 1: Complete a model comparison analysis similar to those discussed in the lectures (the papers from the slides in week 15, for example)
Pick at least 2 (more if you choose) data series to model. These could be different assets or the same asset over different time periods, etc.
– Or –
Pick at least 2 (more if you choose) model specifications to estimate and compare. These could be EGARCH vs. GARCH vs. Unconditional Volatility type comparisons or you could compare two different GARCH specifications, for example
Describe your method for model comparison and describe what you found in the data – patterns, anomalies, robustness, etc.
Option 2: Complete a forecast model with output analysis (see the Forecasting Inflation paper for a good example).
Choose a data series to forecast. This could asset returns, corporate earnings or macroeconomic data such as exchange rates, inflation, GDP, etc.
Describe your method for model specification, output analysis and forecast efficacy evaluation and describe what you found in the data – patterns, anomalies, robustness, etc.
Option 3: A method of your choosing.
Submission: Create 4-7 slides presenting your work. Please make sure to cover the points listed below.
a)Describe the data set that you used for your model
b)Describe the methodology that you chose to select the model specification
c)Describe your output analysis and any model comparison that you did between specifications.
d)Describe your motivation for choosing this data set and/or methodology. What question or process were you investigating?
e)Describe your major conclusions and findings.
0 comments