Pattern Recognition Algorithms (Matlab Implementation)

The common dataset for all machine learning models used here are train_sp2017_v19.csv and result_train.csv for training and test_sp2017_v19.csv for testing.

The following models will be used to classify the dataset into three categories.

1. Bayesian Classifier, Error: 8.8%
2. Ho-Kashyap Iterative Algorithm to place the hyperplane seperating three classes, Error: 18.78%
3. Supervised Learning - K-Nearest Neighbour with K = 3, Error - 11.39%
4. PCA + Bayesian Classifier, Error: 25%
5. SVM (without non-linear kernels), Error: 20%
6. SVM + RBF kernel. Error: 9.7%
7. Unsupervised - Crisp k-means clustering