A Study On The Emotion Tendency Of Swordsman Mobile Game Users Based On Text Analysis

With the development of entertainment functions on mobile phones,mobile games are becoming diversified little by little.Compared with computer games which are difficult to operate and inconvenient,mobile games have satisfied the user’s need to make use of fragmented time for entertainment.And in recent years,the national tide wind makes the swordsman mobile game popular.This thesis conducts an emotional analysis of the online comments of two currently popular swordsman mobile games and realizes to explore the emotional tendencies of users for each property of the games and put forward valuable suggestions to developers at the same time.Swordsman culture has been very popular since ancient times.The positive image of swordsman culture is integrated into the current popular mobile online games,which on the one hand improves the appeal of the games themselves,and on the other hand fits the"chivalrous"experience of users and provides an effective way to carry forward the traditional martial arts culture.Text sentiment analysis is to analyze and conclude the texts which have emotional color and judge the emotional tendency contained in it.The texts processed include online comments on the Internet and online comments containing personal emotional tendencies and subjective feelings.Text content mostly contains emotions such as happiness,anger,sorrow and happiness and subjective attitudes such as praise and criticism.Text sentiment analysis needs to extract these tendencies from the texts for analysis.This thesis selects online comments of Eastward Legend:the Empyrean and A Dreaming RuneScape from the Internet to conduct text sentiment analysis.First,Textrank algorithm is used to select keywords.Second,the improved SO-PMI algorithm based on Good-Turing is used to establish domain-specific sentiment dictionary for research.The experimental results show that the F₁value which means the comprehensive index of accuracy and recall rate of positive words and negative words based on the algorithm is 82.4% and 81.3%,it shows that the accuracy and recall rate of positive words and negative words obtained by the algorithm are both high,which means the result has a good effect.According to the domain-specific emotion dictionary combined with basic emotion dictionary and network emotion dictionary,the final emotion dictionary contains 22,858 positive emotion words and 26,130 negative emotion words.Third,the 60 property nouns with the highest frequency of occurrence of the two games are extracted and artificially divided into 8 categories to form the game property dictionary,and then the comments are classified.Fourth,the degree adverbs,negative words and the symbolic punctuation marks of special sentence patterns are given corresponding weights.The emotional scores of each property review are calculated and the emotional tendencies are judged.The average score of each property is calculated to judge the performance of the two games in each property.Finally,the performance of martial arts mobile game in each property is judged by calculating the total equal score of attributes.The result of research shows that Eastward Legend:the Empyrean is the best in terms of story attributes,but the worst in terms of gameplay,which can be improved by appropriately reducing the weight of the game and increasing the enjoyment of the game.A Dreaming RuneScape performed best in terms of performance and worst in terms of setting,which can be optimized by more reasonable layout of the scene structure,adjusting the modeling size according to the real proportion,and strengthening the authenticity of the light and shadow effects.In general,the performance of swordsman role-playing games in economy is relatively good,but the performance in the scene is not satisfactory,which can be improved by reducing computer domain vulnerabilities in the scene while balancing system optimization and scene optimization.