Research On The Evolution Modeling And Coping Strategies Of Doctor-patient Public Opinion Based On Text Sentiment Analysis

With the comprehensive advancement of the construction of Healthy China and Digital China,my country’s medical and Internet fields are booming.Behind the rapid development has also led to a series of practical problems that need to be solved urgently,such as the unbalanced allocation of medical resources,the frequent occurrence of conflicts between doctors and patients and violent injuries to doctors,the rapid spread of online rumors and online public opinion,and so on.Various problems in the medical environment can easily set off a wave of public opinion on the Internet platform,which in turn triggers the outbreak of public opinion between doctors and patients.How to effectively use the public opinion text data on the Internet platform to analyze the emotional attitude of the public,and explore the evolutionary characteristics and changing laws of public opinion between doctors and patients are the key issues in the current information age.Aiming at the massive data on social media platforms,this paper uses machine learning model algorithms to perform sentiment modeling analysis on public opinion data in complex contexts.By constructing a sentiment classifier in the downstream tasks of the BERT(Bidirectional Encoder Representation from Transformers)pretraining model,and combining it with the LDA(Latent Dirichlet Allocation)topic extraction technology,the semantic information of the topic dimension is increased,and the accuracy of sentiment classification is improved.Combined with the ARIMA(Autoregressive Integrated Moving Average Model)time series model,the simulation prediction of emotional trend can be carried out.Through experimental simulation,the sentiment prediction accuracy of the BERT-LDA multi-dimensional sentiment analysis model of doctor-patient public opinion reached 98%,and the average prediction error of the ARIMA time series prediction model was lower than 11.25%.This fully proves that they can be effectively used in the monitoring and analysis of public opinion of doctors and patients.Then,from the perspectives of region,gender,time,theme,and future,it deeply explores the evolution law and characteristics of public opinion of doctors and patients,and analyzes the distribution of public opinion of doctors and patients in the country from a coarse-grained level.From the fine-grained level integrated with thematic features,the annual evolution of public opinion of doctors and patients is analyzed,and the specific distribution of the focus of public attention,degree of attention-satisfaction in each year,and the moderation performance in the prediction of future doctor-patient relationship are obtained.On this basis,this paper proposes corresponding strategies for responding to public opinion of doctors and patients from the three main bodies of social media,doctors and patients,and government departments,in order to provide scientific and reasonable research and judgment methods for improving the risk warning and response mechanism of public opinions of doctors and patients,which is in line with the public opinion of doctors and patients.realistic management needs.This has important practical significance for building a harmonious doctor-patient relationship,creating a harmonious medical environment,and maintaining social security and stability.