Cricket Match Outcome Prediction Using Tweets and Prediction of the Man of the Match using Social Network Analysis: Case Study Using IPL Data

Abstract

Nowadays, social media has become the main platform that enables people to express their feelings with freedom. In this research, three models were built using features created from tweets and natural parameters. For model creation, machine learning algorithms such as Support Vector Machine, Logistic Regression, Naïve Bayes and Random Forest were used. Final results indicated that the Twitter based model is better than natural parameter-based model. Mixed model was better than both models. Moreover, models were created for every 10 overs using tweets to find how the accuracy of prediction has changed with time. The model created using tweets of the last 10 overs had the highest accuracy. The effectiveness of classifiers were evaluated in each model and Logistic Regression and Support Vector Machine have shown higher accuracy than others. Finally, social network analysis was done to find the most popular player in each match and relationship between being popular and the man of the match was checked.