Classification of Community Complaints Against Public Services on Twitter

Abstract— Information on public services is an important part of increasing community satisfaction with government policies. Complaints and Complaints of the community become mediators to improve public services according to community needs.Twitter is one of the most widely used social media in the community to post activities, experiences, and complaints about public services through the internet easily and realtime.The amount of information on Twitter is mixed between satisfaction and extensibility of public services, making it difficult for the government to make decisions in public policy. The role of Big Data can be a solution to classifying data to predict satisfaction or extensibility of public services with parameters: markets, transportation and hospitals.Data sources taken from Twitter are 700 data texts. The twitter classification of public service complaints is built using the Naïve Bayes Algorithm Method, because the algorithm can classify based on probability values. Text processing is done by filtering text and selecting text to be ordered.The results of this study indicate that the Naïve Bayes Method is able to properly classify public service complaints based on 3 parameters, transportation, markets and hospitals. System testing using 700 data obtained the best results accuracy value: 86%, and precision: 72%, recall 81% and f-measure: 83%