The project is to prepare data and do data mining similar to Time-Series Analysis studied in chapter 11.
You need to pick up one stock to analyze (which is "Groupon").
You will need to download at least one year daily records of Nasdaq, Dow Jones, and the stock you are assigned which is groupon.
You need to prepare your own data file. At least, you need to have today-1 and today-2 in your input.
It’s OK that you have more than two indexes and one stock as input. It’s OK that you include more than two days of historic data.
Pick one algorithm to analyze your data.
Your report should have, but not limited to, the following chapters.
Title Page: Names of your project
Table of Contents.
Introduction: Discuss or give overview of your project and data set.
Goal: Describe the goal of your data mining.
Briefly discuss the algorithms you used in the project.
You should try multiple models (one or more algorithms). You may need to apply other skills, such as, data transformation, attribute selections, etc.
If accuracy is too low, you need to try other algorithms.
If accuracy is too high, you need to make sure there is no over fitting issue.
Do NOT include all data/screen from Weka. For every model, copying/pasting its first and last pages is enough. Do NOT include data records in your report.
Evaluation: Use statistic data or another algorithm to evaluate your model. This chapter is important. Don’t skip it.
Use your product to predict stock price of next day from the last day of your data. What is your error? Will you use your model in decision making for your stock investment?
Product: You need to show usable product. For example, show if-then rules for decision tree and production rules, equation for linear regression, or neural network diagram and equations for neural network.
Summary or Conclusion: Make summary or conclusion for your report.