Data analyzing and processing

Features Correlation Analysis

Feature transformation

Evaluation measures haven been applied

Splitting data into(training - validation - test)

Hyperparameter tunning

• To choose the most fitted hyperparameters for testing

Dealing with imbalanced classes

Mutiple classifier have been testes (to comprehend how each classifiers perform better on a particular sections):

• used classifier include:
  • KNN (KD-trees due to reasonable number of features)
  • Linear regression
  • Decision Trees
  • Data ensembling:
    • Random Forest
    • Bagging
    • Ada boost

Error Analysis

Test predictions