In the context of information overload and the widespread use of social media, the proliferation of fake news poses significant challenges to the credibility of digital information services. This study aims to enhance information credibility assessment by integrating artificial intelligence (AI) techniques with user behavioural and psychological features.
This study was conducted on a sub-set of the Weibo data set. Firstly, the authors used the BERT (bidirectional encoder representations from transformers) model to extract semantic features from news content and user-generated comments. Subsequently, personality traits were inferred from the comment texts using a separately trained BERT-based personality prediction model, which identifies users’ Big Five personality dimensions. Finally, the extracted news features, comment features and inferred personality traits were integrated to improve the performance of fake news detection systems.
Experimental results indicate that incorporating personality traits improves both the accuracy (+1.96%, 90.76%) and F1 score (+1.51%, 90.60%) of misinformation detection models. When considering the five categories of personality traits, conscientiousness contributes most significantly to classification accuracy, whereas neuroticism has the least impact.
This research presents empirical evidence of integrating AI techniques and user behaviour analysis to improve the quality and reliability of digital information services. It offers a novel perspective on using psychological profiling to enhance the effectiveness of misinformation detection.
