1. K BOT - korayozgurbuz/SWE573TermProject GitHub Wiki

This project is created as a part of the course SWE573: Software Development Practice, instructed by Dr. Suzan Üsküdarlı at Boğaziçi University in Spring 2018. The repositories will be written in Python and requirements, user stories, use case diagrams, work schedule and any other documents related to project will be posted on GitHub on a weekly basis.

K-BOT: A Twitter Bot Detection Tool (Spring, 2018)

Twitter bot can be defined as a program which, in general, used to post automated tweets, send auto-replies to tweets having specific words, follow Twitter users automatically or serve as spam to take the attention of the users for fake news on the Twitter.

There are many different types of Twitter bots and their purposes vary from one to another. Although, some bots may tweet helpful material for the sake of humanity, majority of these accounts are, most of the time, undesirable.

How to Spot a Social Bot on Twitter

Detecting non-human Twitter users has been of interests to academics for recent years. There are some projects developed by Indiana University or MIT, and these projects are developed by using machine learning algorithms.

I will develop K-BOT as follows: First, a keyword will be searched and the last 20 users who posted tweets with that keyword will be recorded. After that, I will check at least two conditions. One of the conditions is to check the number of tweets sent by these accounts in a day. If it is more than 50, then the account will be considered as a bot account. Second condition is to check number of retweets of the last 100 tweets of these accounts. If at least 95 of them are retweets, then it will again be counted as a bot account. Currently, I do not have any machine learning algorithm for these project, but i will use Django Framework in Python to configure things and annotate the code to get the functionality I need.