Simple SPAM filter using Naive Bayes classification
Budget: $30 – $250 USD
A spam filtering dictionary is created with 40 entries. 20 examples of spam and ham were compared, yielding a length-40 Boolean vector for each sample. Dataset containing 20 entries for spam and ham are provided for training, and another one with 1000 entries for testing and evaluating.
A python program to do the following:
a) Loading the training and testing data from provided files.
b) computing the minimal sufficient statistics from the training data needed to estimate the classifier parameters. Compute the minimum sufficient statistics needed to classify the data.
c) Compute the classifier parameter estimates in two ways: 1) Max Likelihood estimation and 2) Max A Posteriori Estimation. Provide side by side plots of Ham and Spam parameters estimates. Display the 40 entry hit probabilities for MLE in red points and MAP in green points.
d) Compute the logarithm of the ratio of the A posteriori probabilities (LAPPR) of Spam vs Ham for all testing data.
A python program to do the following:
a) Loading the training and testing data from provided files.
b) computing the minimal sufficient statistics from the training data needed to estimate the classifier parameters. Compute the minimum sufficient statistics needed to classify the data.
c) Compute the classifier parameter estimates in two ways: 1) Max Likelihood estimation and 2) Max A Posteriori Estimation. Provide side by side plots of Ham and Spam parameters estimates. Display the 40 entry hit probabilities for MLE in red points and MAP in green points.
d) Compute the logarithm of the ratio of the A posteriori probabilities (LAPPR) of Spam vs Ham for all testing data.