Description

Title: Implementation of simplified SMO algorithm
References: “ My video lectures on ML: 30-39”.
Full marks-100
Objective: To learn how to train and predict a soft margin-SVM with RBF kernel using SMO algorithm.

1. Use the dataset of heart disease provided on my assignment folder of the course with the following pre-processing and instructions:
• Use only two features for simplicity- age ( data in column #1) and trestbps (on admission to the hospital, data in column #4, i.e resting blood pressure in mm/Hg)
• Modify the last column (# 14) from 1 –heart disease & 0 –no heart disease to Y(i)= {1 and -1}.
• Apply feature scaling methods to the data of Col# 1 and Col# 4.
• Use 70% data for training and 30% for testing.

2. Study the J Platt’s paper on SMO algorithm provided in this folder for your convenience and implement.
4. Always put proper references for the materials you are using.
https://github.com/apex51/SVM-and-sequential-minimal-optimization

Degree of difficulty -7 ( 10 being most challenging)