Logistic Regression Algorithm with One-VS-Rest strategy: in-sample prediction accuracy measure by micro-average AUC is: 98%, out-of-sample prediction accuracy is about 90%. Regularization strength parameter is set as 1e-5, which may be a little small and have caused over-fitting. So next step, in addition to trying other algorithms, tuning parameters to find optimal values will be attempt.

SVM Algorithm with One-VS-Rest strategy: (coming soon)

KNN Algorithm: (coming soon)

Decision Trees and Random Forest Algorithms: (coming soon)