Multi Class support vector machines Weston and Watkins approach

This is an implementation for the algorithm (both training and testing) fully vectorized in numpy (CPU) only, in a shallow neural network paper link : https://www.academia.edu/813112/Support_vector_machines_for_multi_class_pattern_recognition

Pre-requisites

• numpy
• matplotlib

for data handling and preprocessing

• pandas
• sklearn

Getting Started

Preprocess the data doing some dimensionality reduction then tain the model.

Datasets used

CIFAR-100 wheat_seeds