融合多特征的在线社交网络用户人格预测研究  

Research on Personality Prediction of Online Social Network Users Incorporating Multiple Features

作　　者：赵敬华[1] 李澳 张景彦 ZHAO Jinghua;LI Ao;ZHANG Jingyan(School of Business,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《小型微型计算机系统》2025年第2期321-329,共9页Journal of Chinese Computer Systems

基　　金：上海市教育科学研究项目(C2023292)资助;国家自然科学基金青年项目(72201173)资助。

摘　　要：在线社交网络用户的人格与其行为模式、需求偏好和心理健康密切相关.本文提出了一种融合多特征的在线社交网络用户人格预测方法.该方法构建了一个在线社交网络用户人格特征体系,通过提取数字信息中的文本特征、行为特征、语言特征和情感特征,采用早期特征融合策略,使用微博平台数据,对用户的Myers-Briggs Type Indicator(MBTI)人格类型进行预测.实验证明,本文提出的融合多特征的方法相较于简单基于文本特征的方法在分类的效果上更为出色,准确率和F1值分别提升了2.44%、2.59%.同时也表明,CNN在融合多特征的在线社交网络用户人格预测任务上展现出卓越的性能,而BERT结合BiLSTM在人格的信息收集方式和决策方式维度上表现出明显优势.The personality of online social network users is closely related to their behavior patterns,needs preferences and mental health.This paper proposes a personality prediction method for online social network users based on multi-feature fusion.This method constructs an online social network user personality feature system.By extracting text features,behavioral features,linguistic features and emotional features in digital information,it adopts an early feature fusion strategy and uses Weibo platform data to analyze the Myers-Briggs Type Indicator(MBTI)personality type of users.Experiments have proven that the method based on multi-feature fusion proposed in this article is more effective in classification than the method based on text features,with the accuracy and F1 value increased by 2.44%and 2.59%respectively.It also shows that CNN has shown excellent performance in the task of predicting the personality of online social network users based on multi-feature fusion,while BERT combined with BiLSTM has shown obvious advantages in the dimensions of personality information collection and decision-making methods.

关 键 词：人格预测 在线社交网络 多特征融合 MBTI 机器学习 深度学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]