The Addition Symptoms Parameter on Sentiment Analysis to Measure Public Health Concerns

Abstract

Information about public health has a very important role not only for health practitioners, but also for goverment. The importance of health information can also affect the emotional changes that occur in the community, especially if there is news about the spread of infectious disease (epidemic) in particular area at the time, such as case of outbreaks Ebola disease or Mers in specific area. Based on data obtained from Semiocast, Indonesia is the country with fifth largest number of Twitter users in the world, where every topic that lively discussed will also influence a global trending topic. This paper will discuss the measurement of public health concern (Degree of Concern) level by using sentiment analysis classification on the twitter status. Sentiment data of the tweets were analyzed and given some value by using a scoring method. The scoring method equation (Kumar A. et al., 2012) will be tested with new additional parameters, ie symptoms parameters. The value of any twitter user sentiment is determined based on adjectives, verbs, and adverbs that contained in the sentence. The method that we used to find the semantic value of adjectives is corpus-based method. While for finding the semantic value of the verb and adverb we used a dictionary-based method.