关键词:
人格识别
社交媒体
人工智能
特征提取
模型预测
摘要:
人格是与人类思维、情感和行为相关的稳定模式,能够有效帮助我们理解、分析和预测人类行为,被广泛应用于人机交互、推荐系统和网络安全等领域。社交媒体为人格识别研究提供了丰富的数据来源,推动了自动人格识别的发展。然而,现有文献在特征提取与模型预测研究方面,尤其是多样化组合方式的讨论上仍显不足。且人工智能在该领域的应用尚处于发展阶段,亟需进一步探索。因此,本研究综述了面向社交媒体文本的人格识别研究,针对不同任务需求和数据特征,对人格识别模型的组合进行了梳理和总结,涵盖基于语言和统计特征、预训练语言模型特征,以及机器学习、深度学习和集成学习等方法。通过比较分析,研究旨在为从文本数据中推断人格特质提供更有效的方法和策略,推动其在实际应用中的发展。同时,探讨了新兴AI技术在人格识别中的潜力,指出了该领域的研究不足并展望未来方向。Personality, a stable pattern of thoughts, emotions, and behaviors, plays a crucial role in predicting human behavior. Its recognition has broad applications in fields like human-computer interaction and cybersecurity. Social media, with abundant user-generated data, drives automated personality recognition;However, comprehensive studies on feature extraction and model prediction, especially those exploring diverse methodological approaches, remain limited. While AI shows great promise in advancing personality recognition, its application is still in the early stages and requires further investigation. This study reviews existing research on personality recognition from social media texts, focusing on strategies involving language features, pre-trained models, and various machine learning techniques. It aims to improve personality prediction, identify research gaps, and explore future AI-driven developments in the field.