In this project I implemented decision tree based on information gain, k-nearest neighbors, multiple SVM kernels. These learning algorithms on the postal dataset and a document classification dataset
In this project boosting, Online learning (Perceptron) and reduction algorithms were implemented.
In this project, Clustering with K-means and Dimensionality reduction with PCA where implemented.
This project was basically a repeat of previous project’s K-means
implementation, but this time in terms of Gaussian
mixture models.
In this project I implemented the below cascade architecture based classifier for finding synapses in Rabbit retina images based on ICCV'03 paper Learning a Rare Event Detection Cascade by Direct Feature Selection, JianxinWu, James M. Rehg, Matthew D. Mullin.
It was a needle in a haystack problem. The classification wasn't having required accuracy because the features didn't provide required separability for classification.