Deliverables:
—- Project report: data introduction, method used, and final results. 1-3 pages, do not exceed 3, concise is better.
—- Code and data shall be available and ready to run.
Requirements:
—- Individual-based, no group project.
—- Collect multivariate data, where n>100 and p>5.
—- Conduct the following analysis:
Least square using all variables;
model selection (say stepwise) to find out the best possible regression model.
—- Proper evaluation metric:
Using cross validation, or proper training / testing split;
Report MSE (or other proper metric) on testing data.
Grading is based on:
—- Whether your code runs properly, and outputs matching the slides.
—- Whether the method and evaluation criterion is properly used.