Method for classifying usability qualities and problems for action games from user reviews using text mining

Abstract

Game reviewing is one of the method for game users and critics to comment and discuss about a game. Game developers and marketers could use game reviews as insights to assist on designing a better game by specifying quality requirements and providing better game marketing. Usability and problems are major concerns of users and game developers since these quality affects users' satisfaction and opportunity to increase market share respectively. However, manually extracting game usability qualities and problems from a large number of game reviews is challenging since most reviews often use natural language and may not be well-organized. This research presents a method for extracting and classifying game usability qualities and problems from game reviews using text mining technique. This research is composed of 2 main phases. The first phase is specifying game usability qualities and problems categories which also included keyword corpus creation for each category. The second phase is gathering, extracting, analyzing and classifying game reviews. The gathered game reviews will be extracted and classified into each usability qualities and problems category. Game usability qualities classification will also include sentiment classification to further specify positivity and negativity of each sentence. Accuracy metric will be used for classification evaluation. In addition, a game review summary on usability qualities and problems will be created for developers to improve game quality.