Hi! I'm working on an exercise about Spam Filters using Naive Bayes.

Got the data from TREC06 Public Spam Corpus. However, I'm having trouble starting.

Basically there is a file that contains directory of emails in another folder. Format is as follows:

ham ../data/000/000 spam ../data/000/001 spam ../data/000/002 ham ../data/000/003 spam ../data/000/004 ham ../data/000/005 ham ../data/000/006 spam ../data/000/007

This file basically says that in the folder data > 000 > 000 is ham. Then data > 000 > 001 is spam, and so forth.

The folder named data has all the emails (1 email = 1 file).

Wondering how I can create a dataset that has all the email body in one? Kindly help me on the python code please.

Index	Classification	Email content
000/001	Ham	This is the email body
000/002	Spam	This is another.