graydonhope / Spam-Classifier

Using Support Vector Machines to build a Spam Classifier

Spam Classifier using Support Vector Machines

Using an SVM with Gaussian Kernels to do non-linear classification

Gaussian Kernel Function:

Finding non-linear decision boundaries

In effect:

Training an SVM for Spam Classification

After the emails are pre-processed, the program extracts features from the email and converts them into a vector in R^n

After training on a dataset from the SpamAssasin Public Corpus, the classifier detects

98.5%

of spam emails correctly.

Course base on Stanford's Machine Learning Course.