Design And Implementation Of Post-medical Evaluation System Based On Deep Learning

One of the most important reasons for the current tension between doctors and patients is that there is too little effective communication between doctors and patients,and the online evaluation system is an effective communication channel between doctors and patients.In the long run,it is helpful for medical institutions and doctors to improve service quality and gradually alleviate the conflicts between doctors and patients.It can provide a reference for the patient to choose a doctor for medical treatment,provide channels for the doctor to listen to the patient's opinions on the diagnosis and treatment,it can also help doctors improve the work according to the patient's evaluation and needs,and the medical institution can also give appropriate rewards and punishments based on the patient's evaluation of the doctor to improve the overall management level of the hospital.However,a doctor may have hundreds of evaluations,in a medical institution,hundreds of doctors will have a large amount of information on patients' evaluation texts,and it takes a lot of manpower and time to analyze and identify positive and negative evaluations by only manual methods.At the same time,a large amount of evaluation data cannot provide effective and concise information for patients,doctors,medical institutions,etc.How to extract valuable core information from it is also an urgent issue to be solved.The main task of this paper is to design and develop a post-medical evaluation system based on deep learning for a tertiary hospital,and this system is a subproject in the medical integration project of the hospital.This paper combines the existing problems with the actual needs of the hospital,designs and develops the post-medical evaluation system based on the MVC architecture,realizes the basic functions of the online reviews,and proposes a sentiment analysis method for sentiment analysis of word-oriented,applies keyword extraction techniques to review the comment text data analysis,integrates the review information resources,accelerates the construction of medical information in the hospital,and at the same time,it provides services for users on the web and mobile terminals.The details are as follows:1)Designs and implements a post-medical evaluation system.This paper adopts MVC design mode to build this post-medical evaluation system,and provides services for patients,doctors,and administrators.On the web side,we implement basic functions such as user reviews and doctorpatient communication,and provide evaluation-related services for patients on the mobile terminal,improving users' experience and system flexibility;we use the character-oriented method for sentiment analysis in the system to implement the function of sentiment analysis on medical comments;combining the keyword extraction of comment data to achieve the function of comments analysis and display.The layered idea adopted by the system reduces the degree of coupling between the modules,and makes the logic design of the business code clearer.It is convenient for the later maintenance and modification,making the information exchange of each functional module easier and more convenient.2)A character-oriented method for sentiment analysis.The method first uses the pre-trained word vector to compose the characteristic matrix of the sentence as the input of the convolutional neural network,it eliminates the extra word manipulation and text noise caused by lexical analysis and word segmentation errors for the word vector.It uses the back propagation algorithm to train the model,then uses the gradient descent method to optimize the parameters.In the training process of the model,when the pooling operation is performed,the text is divided into different parts by punctuation and other delimiters,each part uses the pooling layer to maximize the feature value to retaining multiple eigenvalues in the sentence,and it can better to capture the emotional information contained in the sentence.3)Implements keyword extraction for comment text.In the data analysis of the reviews,keyword extraction technology is introduced to classify a large number of review data of doctors and extract the core information of the patients' evaluation of doctors,solve the problem of information overload when users face massive information.According to the extracted keywords,a specific characteristic cloud image is generated for each doctor.The cloud map will represent the weight of each keyword based on the size of the keyword,and the patients,the doctors,the administrators are provided with a visual display of the patients' description of the doctor's characteristics.