Application And Research Of Emotion Classification Technology Based On Deep Learning For MSM Emotion Classification

In recent years,the number of AIDS patients has been increasing,and studies have found that transmission through MSM(Men who have sex with men)is an important way of transmission.In order to prevent the spread,a simple and quick identification method is needed to judge whether MSM is sick or not.At present,the early warning research on the condition of the target person's emotion has become a hot spot in the field of medicine and science and technology.However,due to the simple content of the dialogue content of the target person,the amount of data is complicated,and the information judgment efficiency is low,according to the semantics is not enough to support the rapid judgment of the disease,a quick classification judgment method is needed to warn the disease in order to find the patient in time Block.This article is a technical application of sentiment classification in the judgment of illness.Based on the accurate classification of text sentiment,the target person's AIDS prevalence is quickly identified.In order to improve the early warning efficiency of the disease,alleviate the problems of slow recognition speed and low accuracy.This paper combines the sentiment classification algorithm and the text mining algorithm to improve the traditional BERT(Pre-training of Deep Bidirectional Transformers for Language Understanding)algorithm,and proposes a hybrid BERT algorithm.The algorithm improves the fully connected layer of BERT,finds its optimal threshold to improve accuracy,and combines the idea of text mining KNN(K-Nearest Neighbor)to perform corresponding weight calculation on text sentiment and keyword informationto achieve The purpose of multiple extraction of text information,so as to realize disease detection and early warning based on text information.In the experimental stage,through the use of Python-based hybrid BERT classification early warning algorithm and other algorithms to compare experiments to verify the superiority of the hybrid BERT algorithm proposed in this paper.The main research content of this article is divided into three parts: first,research on the emotion classification process,and focus on the description of the classification method;second,research on the emotion classification technology in the judgment of the disease,from deep learning and text mining based Discuss and discuss two processing methods;third,research and discuss the hybrid BERT algorithm proposed in this paper,which inherits the accuracy and efficiency of BERT and has the ability to classify the condition based on emotion,so it is applied to short conversations based on text Judging and warning about AIDS.Experiments show that the hybrid BERT algorithm proposed in this paper can timely judge the prevalence of target people through text analysis.This way of judging and warning the AIDS condition through emotion classification technology has certain reference application value in the field of medical diagnosis.