The Application Of Deep Learning In Semantic Sentiment Analysis

With the improvement of the level of intelligence in life,people are no longer limited to the single way of expressing their views on things in the past,and people can freely express their opinions in multiple ways and through multiple channels.The processing of massive information has gradually eliminated the traditional data processing technology,and how to quickly and accurately extract the emotional information in the massive information needs to be solved urgently.Semantic sentiment analysis technology in natural language processing can extract explicit or implicit emotional expressions from massive subjective information collected from various channels.Based on this,the practical application of semantic sentiment analysis technology shines brightly.In addition,the sudden appearance of the epidemic has led to closed management in colleges and universities,and the psychological and emotional health of students cannot be ignored.This paper studies the semantic sentiment analysis technology in deep learning,performs algorithm fusion and improvement,and designs and implements a semantic sentiment analysis system for school students.The main tasks are as follows:First,the related technologies such as bidirectional long-term and short-term neural memory network,capsule network and attention mechanism applied in semantic sentiment analysis in deep learning are studied,analyzed and interpreted.Second,the semantic sentiment analysis model BAC in deep learning and its optimization and improvement T-BAC are designed and implemented.(1)The effective combination of the bidirectional long-short-term neural memory network,the capsule network and the attention mechanism is studied,and a semantic sentiment analysis BAC model that can obtain the overall features and high-level features of the data is designed.First,the model extracts contextual time series features through a bidirectional long-term and short-term memory network,and uses the attention mechanism to assign reasonable weights to the output vectors of the front and rear hidden layers to improve the quality of semantic feature expression of the model.Finally,the capsule network aggregates the above semantic sentiment features through dynamic routing to obtain high-level features,thereby improving the overall semantic sentiment analysis effect of the model.(2)The different weight distributions of different temporal semantic features are comprehensively considered,so as to design an optimized and improved T-BAC model.(3)A dataset of semantic sentiment comments of school students is constructed,and the validity of the T-BAC model and the correctness of the model improvement method are verified by experiments with other datasets mentioned in this paper.Third,use C#+My SQL+Python technology to design and implement the semantic sentiment analysis system for school students.Applying the T-BAC model designed in this paper to the system,analyzing the single or batch semantic emotional comments of students in school,can clearly understand the emotional polarity of students in all aspects of the school,and assist school administrators in the next steps.One-stage student management.After functional page display and system testing,the expected design requirements have been achieved.