Multi-modal feature fusion method and system for sentiment analysis

The invention provides a multi-modal feature fusion method and system for sentiment analysis. According to the scheme, the method comprises the steps of obtaining text data in social data and image data corresponding to the text data; performing feature extraction on the text data and the image data to obtain text features and image features; based on the text features and the image features, obtaining interaction information between the text features and the image features by adopting a cross-modal attention mechanism, and performing noise filtering on the interaction information through a gating mechanism to obtain text features after image filtering and image features after text filtering; splicing the obtained text features, the image features, the text features after image filtering and the image features after text filtering to obtain fusion features; and based on the fusion features, obtaining an emotion analysis result through a pre-constructed emotion analysis model. 本公开提供了一种用于情感分析的多模态特征融合方法及系统，所述方案包