Predicting Information Diffusion in Online Social Platforms: A Twitter Case Study

Online social media has become a part of everyday life of modern society. A lot of information is created on these platforms and shared with the community continuously. Predicting information diffusion on online social platforms has been studied in the past by many researchers as it has its applications in various domains such as viral marketing, news propagation etc. Some information spreads faster compared to others depending on topic of interest of the online users. In this work, we investigate the information diffusion problem using Twitter data as a use case study. We define tweet popularity as number of retweets any original message receives. In total we extracted 27 features which can be categorised into content, user, sentiment and initial retweeting behaviour for creating our prediction model. We study the problem of predicting as a multiclass prediction task. Three datasets from Twitter about three different topics are collected and analysed for building and testing various models based on different machine learning algorithms. The models were able to predict up to 60% of overall accuracy and an F1 score of 67% is obtained. The models are created using one of the dataset and tested on all the datasets, which shows that the model is robust enough to handle different topics.