Search

People participation systems for public comments and suggestions on government projects, using web-boards for public participation have a variety of content and the text data cannot be interpreted. The messages are not clearly categorized without summary of comments and suggestions data on government projects and cannot demonstrate public responses. This article aims to present a classification system for public comments towards government projects with artificial intelligence methods. This study used experimental research methodology with the model evaluation. The four indicators included precision value, recall value accuracy value and F-Measure value were measured. The study indicated that the model evaluation details are as follows: three data characteristics are categories of comments and suggestions indicated that precision = 0.88, recall = 0.86, accuracy = 86.69% and F-Measure = 0.87 ; project type indicated that precision = 0.87, recall = 0.85, accuracy = 85.23% and F-Measure = 0.85 ; and government agencies or provinces indicated that precision = 0.86, recall = 0.83, accuracy = 83.94% and F-Measure = 0.84. [ABSTRACT FROM AUTHOR]