Shiyu Mou shiyumou@usc.edu
- Data is ploted as scatterplots and boxplots.
- KNN is used to make classifications. Different parameters were tested.
- k
- Number of training set
- Distance metrics (Euclidean, Manhattan, Mahalanobis, Chebyshev)
- Decisions (majority polling & weighted decision)
- Precision, reall, F-score, Learning Cureve were derivated to understand the performance.