Game Comment Data Mining Based on Text Sentiment Analysis

Jiaxin Qian,Genfu Yang
DOI: https://doi.org/10.1007/978-981-15-1864-5_76
2020-01-01
Abstract:In order to understand the needs of the player, understand the advantages and disadvantages of the game, this thesis proposes a way to evaluate the game by player comment content: (1) from the site to mine comment information. (2) Calculate the importance and score of each attribute through the neural network. (3) Analysis via importance-performance analysis (IPA) using attribute importance and scoring. In this way, the advantages and disadvantages of the game can be clearly demonstrated in the form of charts. This issues the combination of dictionary and machine learning, the text vector of emotional words, negative words, degree adverbs, punctuation and whole sentences is characterized by the emotional analysis of sentences, and the evaluation and analysis of the game in five dimensions by IPA method is used.
What problem does this paper attempt to address?