Sentiment analysis in Twitter about Human Papillomavirus

Abstract

Internet has grown and expanded enough to become the main source for finding any kind of information. Moreover, misinformation and hoaxes blend in with trustful information, and most users are not able to distinguish between them. This becomes dangerous when the information in question is health related. Human papillomavirus (HPV) has become in one of the most common sexual infections around the world and it is particularly dangerous due to the fact it can lead to some types of cancer. Two vaccines are available and taking them is recommended before reaching the onset of sexual activity, from 11 through 12 years old. Nevertheless, the relative novelty of the vaccine and its sexual origin translate into low rates of vaccination and the spread of opinions against the vaccine on the World Wide Web. This thesis presents the design, implementation, testing and evaluation of a sentiment classification system of texts obtained from Twitter. The system is evaluated by checking the statistics of each groups’ opinion and measuring the algorithm precision in comparison with a human manual classification.