User Multi-Modal Emotional Intelligence Analysis Method Based on Deep Learning in Social Network Big Data Environment

Abstract

In order to accurately analyze the emotional tendency of social media users' evaluation and better promote the work of emotional analysis and recommendation algorithm. This paper presents a new text emotion classification model, which integrates content features and user features, by representing the sentence, content features and user features of microblog as vector matrix and inputting them into the text emotion classification model which integrates content features and user features. Firstly, it analyzes the content information and user information related to the sentence emotion of the target microblog, and constructs the content characteristics and user characteristics respectively; Then, a text emotion classification model is constructed, which integrates content features and user features; Then, the method of feature level fusion and decision-making level fusion is designed for microblog emotion analysis of image and text fusion. Maxout neuron is also introduced to solve the problem of gradient dispersion in the training process and optimize the training process. The experimental results show that: Compared with other models, the accuracy of the proposed model is improved by more than 2.5%, and it is better than other models in recall rate and F value.

Keywords

• Multimodal affective analysis
• deep learning
• Maxout neuron
• long and short term memory network (LSTM) model
• intelligent analysis
• feature fusion