Skip to content

In this project, we will use Machine Learning techniques to predict weather an account on Twitter is a Bot or a real user. We have performed significant amount of feature engineering, along with feature extraction - selected features out of 20 helped us to identify whether an account is bot or non bot. We implemented various algorithm but finall…

Notifications You must be signed in to change notification settings

acm-iem/Twitter-Bot

Repository files navigation

Twitter Bot

What it meants to do?

Twitter bot is an automated user account that interacts with Twitter using an application programming interface (API). This bot is programmed to perform tasks normally associated with human interaction such as

  • like tweets,
  • and, most importantly, they can tweet content, and retweet anything posted by a specific set of users or featuring a specific hashtag. (here, #acmiem and also #csiiem)

It is also programmed to produce automated posts, follow Twitter users or serve as spam to entice clicks on the Twitter microblogging service.

Requirements

  • Python 3
  • Tweepy

Procedure:

We need to get the following credentials/API keys from the developer.twitter.com :👇

consumer_key = 'consumer key'
consumer_secret = 'consumer secrets'
access_key = 'access token'
access_secret = 'access token secret'

Dependencies:

Before running it in your system, make sure you install ⬇ all the dependencies below:

npm install twit

npm install socket.io


The twitter real-time stream tracks the the specific hashtag (which Can be customised as well) and the Bot Likes and Retweets the Tweet.🔁

Detecting Bots on Twitter Using Machine Learning

In this project, we will use Machine Learning techniques to predict weather an account on Twitter is a Bot or a real user. We have performed significant amount of feature engineering, along with feature extraction - selected features out of 20 helped us to identify whether an account is bot or non bot. We implemented various algorithm but finally implemented our own which gave us AUC>95%.

Requirements

  • Python 3
  • Pandas
  • Numpy
  • SeaBorn
  • MatplotLib
  • Sklearn

The code and datasets can be found under 'Final Project and Code' folder.

Steps followed

💻 First the data has been collected using Twitter API

💻 Analyzed important features (Pre-processing)

📎Followers vs Following Count

📎Frequecy of appearance of Bots vs NonBots in ScreenName

💻 Used Decision Tree Classifier

Contributing

Contributions are always welcome!

🌸🌸HactoberFest 2021🌸🌸

See contributing.md for ways to get started.

Please adhere to this project's code of conduct.

License

MIT

Reference

https://www.youtube.com/embed/bRjxeovhL50?ecver=2

About

In this project, we will use Machine Learning techniques to predict weather an account on Twitter is a Bot or a real user. We have performed significant amount of feature engineering, along with feature extraction - selected features out of 20 helped us to identify whether an account is bot or non bot. We implemented various algorithm but finall…

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published