Text Extraction And Integration Method For Game Reviews

In recent years,the computer game industry has developed rapidly,which constantly enriches the entertainment life of users and generates huge economic value.However,a thriving market also means a competitive environment,and developers need to keep their games diverse and innovative in order to attract and maintain customers and keep the market vibrant.Most online game stores allow users to review the games they buy,and developers can read player reviews to see what users think about the game.While fixing bugs in the game,it is also necessary to understand the changes required by the user.Therefore,studying game reviews can help game developers understand users' concerns and further improve the perceived quality of their games.At present,information extraction mainly focuses on feature extraction or keyword clustering search methods.However,neither method can accurately extract the information in game reviews.The feature extraction method performs well in reviews when analyzing common applications,except for the game category.On the one hand,the game industry is more professional,and different types of games have different characteristics.There are a lot of difference between regular app reviews and game reviews,and there are a lot of difference between word understanding.On the other hand,the information content of game reviews is large,the plot is rich,the style is different,the scene and the character characteristics all affect the result of information extraction.Existing research fails to address these questions,and this is the kind of analysis that is currently lacking for game reviews.In view of these shortcomings,this paper designs a method of information extraction and integration for game reviews.It helps developers understand what features users want to add,and it also helps them identify problems in the game that can be used to further improve the game.One of the biggest problem is relevant to the developer.The classifier is trained by constructing a set of keywords to classify game reviews and screen out reviews relevant to developers.The type of information in the filtered comments is also varied in order to help developers understand the information in the comments more easily.We relevant reviews based on a set of keywords,and then divided the filtered game reviews into four categories.For these four categories,we set up different information extraction rules based on grammar tree to cluster,sort and visualize the extracted information.Developers can read the comments and make changes in time.It can also better understand users' cognition of the game,so as to improve their game experience.