Font Size: a A A

Research Of Mobile Game Recommendation Based On Sentiment Orientation Analysis Of Chinese Comments

Posted on:2018-11-02Degree:MasterType:Thesis
Country:ChinaCandidate:M T XueFull Text:PDF
GTID:2348330533466786Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Mobile internet has penetrated all aspects of human activities,and mobile phone games has become a major part of human entertainment.The rapid growth of mobile terminals,various game products,powerful function of smart phone,and the faster speed of 4G network have brought extraordinary experience for the old and new mobile game users and promoted the rapid development of mobile game industry.Under such background,it has become a very important research task for how to seize the golden age of mobile games and provide higher quality game recommendations to mobile game users.At present,there are many comments for each mobile game on every game forum,game site and APP store.These comments reflect the user's direct feeling in a certain degree once they play the game.Meanwhile,it has become a very important reference index for new users to select games.Analyze the sentiment tendency for game comments,and dig into the comments and emphasis of game products will contribute to recommend a game which meet the user's real needs.Based on the above project background and research significance,the main goals of this article include:Firstly,propose a new field words filter method based on PMI.Use words vector to recognize and check the polarity of the sentiment words,so as to build the domain dictionary and sentiment dictionary.Secondly,achieve sentiment orientation discrimination based on deep learning methods and sentence vector features,and then,forming a game ranking according to the favorable score of mobile games.Thirdly,design an extraction method which mix grammar rule and shortest path which will dig into and analyze the comment information of mobile games,and extract the corresponding appraisal collocation.The experiment result shows that,the algorithms about out-of-vocabulary detection and sentiment orientation analysis proposed by this article can handle the large scale of review data,realize the intelligent sentiment judgment,and reduce the comment handling cost.Mobile game recommendation algorithm which based on comment collocation extraction will improve mobile game recommendation accuracy.It has a certain reference value for improving existing game design and new game's functional design,and help game users to find a mobile game meet their own needs.
Keywords/Search Tags:Sentiment Orientation Analysis, Mobile Game Recommendation, Comment Collocation Extraction, OOV(out-of-vocabulary) Detection, Game Ranking
PDF Full Text Request
Related items