Design And Implementation Of Medical Recommendation System Based On Multimodal Cardiovascular Disease Data

In recent years,with the development of Internet technology and Intelligent medicine,the way users get medical information also changes,and they begin to get medical information through the Internet.Due to the accumulation of a large number of multimodal medical information data in the medical field and Internet,the problem of information overload appears,and it is difficult for users to get the relevant medical information quickly.Recommendation system is a common solution to solve the problem of information overload.Medical recommendation system can refine and apply these data,accurately recommend medical information for patients,and optimize the allocation of medical resources.Based on the medical data of cardiovascular disease,combined with clinical text modal data and image waveform modal data,this paper uses hybrid recommendation method to build personalized medical recommendation system to provide recommendation service.The main work of this paper is as follows:Firstly,in view of the problem that the medical recommendation system is not fully utilized for multimodal medical data,this paper proposes a flow method for the acquisition,processing and feature extraction of multimodal cardiovascular disease medical data combining medical text data and image waveform data.Firstly,the multimodal data is obtained and preprocessed,and then the TF-IDF(Term Frequency & Inverse Document Frequency)algorithm is used to obtain the features of medical text data.Then,in view of the difficulty of acquiring structural features of ECG signals in the image waveform modal data,this paper uses the deterministic learning theory to model ECG signals,obtain the Cardiodynamicsgram(CDG)and calculate the quantitative features.Finally,the text modal features and ECG features are combined into multimodal features for storage,which can increase the feature expression ability of the recommendation system.Then,based on the multimodal features obtained above,a hybrid recommendation algorithm model is designed.The model is based on Deep Factorization Machine(Deep FM)algorithm and traditional recommendation algorithm,and is constructed by Parallelized Hybridization.The hybrid model solves the problem that traditional recommendation algorithm models lack the ability to express multimodal data.In addition,the model is compared and verified on the self-built data set.In the experiment,the F1 value(the harmonic average value of precision and recall)of the proposed model reaches 82.51%,which proves the effectiveness of the recommendation model.Finally,on the basis of the above,this paper proposes a medical recommendation system framework of multimodal cardiovascular medical data,and constructs a personalized medical recommendation system based on Browser/Server(B/S)and Flask framework.From the perspective of interactive use,the system realizes the personalized medical information recommendation function.In actual use,by inputting personal multimodal medical information,users can obtain personalized medical information related to cardiovascular diseases such as recommended hospitals,doctors and possible diseases.This design helps to improve the system of tiered medical services and optimize the allocation of medical resources.