Sentiment mining of community development program evaluation based on social media

Abstract

In community development programs, evaluation to determine the achievement level of success and the previous study conducted quantitative and qualitative analysis. However, there is another problems that the number of participants in each activity disable toconclude these activities has high success rate. This study is search the sentiment of every participant in activities based on a tweet from social media. Thus, we focus on quantitative analysis cope with mining sentiment in every tweet related activities. First, we did preprocessing, reduction feature using principle of component analysis and estimation parameter c of classify algorithm. Second, we were modeled sentimentof activities using support vector machine. The last,we performedby calculating term score using term frequency which is combined with term sentiment scoresbased on lexicon.The results shows that models can be summarized sentiment that point out indicate the success level by the mostpart in amount of positive sentiment.