Research On Smart Marketing Of Medical Electronic Commerce Based On Multi-model Fusion For Disease Prediction

In recent years,both the improvement of national health awareness and the disease prevention service objectively post the higher realistic requirements for the accuracy of disease prediction and the timeliness of medicine,medical electronic commerce also has received great attention from the nation.At the same time,although smart marketing,which based on big data,has been applied to more and more situations.But the current research on smart marketing of medical electronic commerce is still inadequate,and many researches ignore the importance of external data,and these researches have the problem of information lag and the inability to reflect the real needs of the market in advance.For medical electronic commerce,the disease prediction is highly relevant and forward-looking external data.Disease prediction can comprehensively and timely reflect the overall requirements of medicine.Therefore,how to implement smart marketing of medical electronic commerce combined with disease prediction has become a new direction of research.In view of these,this dissertation takes the disease prediction's outcomes as external supplemental information of the medical electronic commerce.On this basis,firstly,this dissertation proposes a disease prediction model based on multi-model fusion,it fuses ARIMA time series model which uses disease history values as the object of regression,and XGBoost(eXtreme Gradient Boosting)model which uses the network search data from baidu as its features.Both the Gaussian transformation of the target variables and multi-model fusion form the disease prediction model.Then this dissertation proves that the model can effectively improve the accuracy and stability of disease prediction.It also proves that the disease prediction is feasible and credible.Secondly,this dissertation explores the content of smart marketing of medical electronic commerce which is based on disease prediction.Thirdly,it builds a smart marketing system of medical electronic commerce which is based on disease prediction.In this dissertation,the disease prediction model is optimized.The data processing method of applying Gaussian transformation to target variables can effectively improve the effect of model fitting.At the same time,this dissertation proposes a disease prediction model,which is based on multi-model fusion,to improve the accuracy of disease prediction.And this dissertation applies the disease prediction to the smart marketing of medical electronic commerce,which provides a guide for the realization of medical electronic commerce's smart marketing.This dissertation has positive practical significance and theoretical value for promoting medical electronic commerce to achieve the information integration and the resource sharing and the improvement of marketing efficiency and the realization of smart marketing.