Please, read wiki for more information

The framework is written mostly in Python 2.7.

Dependencies: scikit-learn, Lasagne, NLTK, matplotlib, sqlite3, GeoPy, pyenchant, igraph

The framework allows to predict human behavior in the context of social networks and media. For example, the framework allows to predict when a user will get involved in an interaction on a website or make an action.

The framework provides an API and has been tested on Twitter, Facebook, and Stack Exchange forums. For example:

• we predicted when a user will answer a question on a Q&A forum;
• we estimated if a user will reply or retweet a message on Twitter;
• we found who the most probably will 'like' a post on Facebook.

The framework allows to explore communication on social networks, analyze the data, choose a prediction task, choose features for prediction, machine learning (ML) algorithm. Then a user can explore the results of the prediction, its accuracy and time performance.

Feel free to use the code under MIT license.