The project aims at developing a spam detector. The algorithm uses supervised classification to achieve this. The training data used is the UCI Spambase data set and the code has been written in Octave. Two Training Models have been implemented (as suggested by the names of the 2 folders in the root directory): 1. naive bayes 2. full bayes (or the decision boundary based bayes classifier assuming normal distribution of the features) *both models assume independence of the features. In both implementations, there are 3 main files: 1. buildModel.m : builds the model and stores it into files. 2. spamDetect.m : loads the model from the files, reads the test data from the specified file and produces evaluation results in the form of accuracy, precision, recall etc 3. spamEval.m : performs n-fold cross validation of the model