The Role Of Pre-Processing On Unstructured And Informal Text In Diabetic Drug Related Twitter Data

摘要

The Ubiquity of Online Social Networks(OSNs) is creating new sources for healthcare information, particularly in the context of pharmaceuticaldrugs. Opinion mining on twitter data is an emerging topic in research. Tweets are usually short, more ambiguous and contain a huge amount of noisydata. Sometimes, it is difficult to understand the user‘s opinion. The first step of the opinion mining is text preprocessing of Twitter data. This researchpaper focuses on the preprocessing techniques to enhance the accuracy of the opinion classification. Twitter‘s contents include users‘ behaviors, statesof mind, comments on certain topics etc, and a lot of these contents express the users‘ opinions unavoidably. Data Preprocessing is the process used toclean useless text from unstructured text for further analysis. The preprocessing is the most difficult task; since it can be done in various methods appliedin twitter dataset. The present paper is based on the demonstration of a complete step-by-step process of analyzing opinions from tweets related tosome specific diabetic drugs. R-tool is used for performing all essential steps. This research work proves as an initiative process to identify diabeticdrugs(generic and brand) and extract potential adverse effects by analyzing the content of twitter messages using opinion mining analysis.