Reference no: EM13103363

Project Title: Email Spam Filter

Abstract:

Analyze the emails and predict whether the mail is a spam or not a spam.

To work on this,

Create a training file and copy the text of several mails and spams in to it And create a test set identical to the training set but with different examples.

For this, you have to implement a grammatical structure by first tagging the part-of-speech (e.g. lots of spam is usually random garbage populated with keywords, where non-spam usually makes some sense). So that you can potentially implement several different classifiers, and compare their accuracy.

Suggestions

First concentrate on using the Naive Bayes classifier. During the presentation, re-explain the classifier and its Bag-of-Words assumption.

It may be possible to extract _your own_ mail data. You can compute unigram probabilities whether an email is spam or not.

1. download all email that is not in a spam folder
2. download all email from your spam folder
3. process this data into nltk

A web interface that accepts an email and outputs either SPAM or NOT SPAM, would make a final piece of the project.

• The most important aspect of the project is the quality of the work.
• a presentation which summarizes your topic
• a demo of a specific NLP system or tool
• a code walkthrough of something you have implemented
• or a hands-on lab which gives experience with a specific technique.

The write-up should be 4-6 pages and describe descriptions of algorithms, resources used, and all results. Implementation, algorithms, data, and analysis of results are all important. Include any supplemental files including your overheads, lab exercises, or other materials.