Machine-Learning-CS60050

All projects (assignments) undertaken as part of my Machine Learning Course (CS60050)

Vehicle Insurance Prediction

• Build Decision Tree Classifer using ID3 algorithm with reduced error pruning using information gain measure.

• Train Naive Bayes Classifier after outlier removal and calculate accuracy with 10-fold cross validation.

• Apply laplace correction to get final accuracy.

Lung Cancer Analysis

• Apply Principal Component Analysis for feature selection (preserve 95% of variance)

• Apply K-means clustering over varied k values(2-8) and report value for which NMI (Normalised Mutual Information) is maximum.

• Apply binary SVM classifier with differnet kernels to obtain best accuracy over normalised dataset.

• Train Multilayer Perceptron models with varying density and hyperparameters. Tune the learning rate to get best accuracy.

• Use forward sleection method for feature selection and apply ensemble learning (max-voting) to get max accuracy.