Modelling and Sentiment Analysis of Game Reviews

游戏评论建模与情感分析

• 批准号: 2743745
• 金额: --
• 依托单位: Queen Mary University of London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2022
• 资助国家: 英国
• 起止时间: 2022 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2743745
• 关键词: Modelling; Sentiment; Analysis; Game; Reviews

Nowadays, companies have access to an unprecedented amount of data from their users. However, extracting the knowledge that these companies need has proven to be a challenging task, and the gaming industry is not an exception. Traditional classification algorithms, such as Logistic regression and Naïve Bayes, as well as more recent ones that use (Deep) Neural Networks, such as Recurrent Neural Networks (RNNs) based algorithms, and the SotA algorithms, such as attention models, are three classes of algorithms that are frequently used for such applications. While the traditional algorithms are the most interpretable, the latter two have superior performance. During my PhD, I aim to investigate each of these algorithms for modelling and sentiment analysis of game reviews on social media. These results are aimed to be used by the game industry to better analyse their users' feedback and utilise various NLP tools to enhance their user experience.

如今，公司可以从用户那里获得前所未有的数据量。然而，提取这些公司所需的知识已被证明是一项具有挑战性的任务，游戏行业也不例外。传统的分类算法，如逻辑回归和朴素贝叶斯，以及最近使用（深度）神经网络的算法，如基于递归神经网络（RNN）的算法，以及SotA算法，如注意力模型，是经常用于此类应用的三类算法。传统的算法是最容易解释的，后两种算法具有优越的上级性能。在我的博士学位期间，我的目标是研究这些算法中的每一个，用于对社交媒体上的游戏评论进行建模和情感分析。这些结果旨在被游戏行业用来更好地分析用户的反馈，并利用各种NLP工具来增强用户体验。