Yendall / YouTube-Spam-Classifier

Machine Learning methods to classify YouTube comment spam

YouTube Spam Classification

Author: Max Yendall

Table of Contents

1. [Explanation of Files] - Explanation of all included files in this package
2. [Running The Scripts]
   • [Running main.py]

Content

1. [Explanation of Files]:

   • main.py: The main module modelled off the given template. Simply run this to step through all exploration, parsing and classification

   • include/ClassificationModule.py: The main classification class which classifies using Naive Bayes and K-Nearest Neighbours using Pipelining (sklearn)

   • include/ExplorationModule.py: The main exploration class which explores statistical summaries and visualises relationships between variables

   • include/DocumentCollection.py: The main parsing class which filters the data and creates a collection of spam documents

   • include/SupportVectorMachine.py: The main SVM class. Includes feature extraction and classification using a Linear Kernel Support Vector Machine

   • include/Settings.py: A settings file to set up the environment, including NLTK redirection and root folders for data and project referencing

2. [Running the scripts]:

   In order to run all four parts of this project, simply the following command from an iPython environment:

   %run main.py