Indian Journal of Science and Technology

Abstract

A social media is a mediator for communication among people. It allows user to exchange information in a useful way. Twitter is one of the most popular social networking services, where the user can post and read the tweet messages. The tweet messages are helpful for biomedical, research and health care fields. The data are extracted from the Twitter. The Twitter data cannot classify directly since it has noisy information. This noisy information is removed by preprocessing. The plain text is classified into health and non-health data using CART algorithm. The performance of classification is analyzed using precision, error rate and accuracy. The result is compared with the Naïve Bayesian and the proposed method yields high performance result than the Naïve Bayesian. It performs well with the large data set and it is simple and effective. It yields high classification accuracy and the resulting data could be used for further mining.
Keywords: CART, Classification, Decision Tree, Machine Learning, Twitter