Design And Implementation Of Tumor Knowledge Recommendation System Based On Hybrid Prediction Algorithm

As the population ages,the number of new cancers and deaths in my country continues to rise,leading most people to believe that suffering from cancer is equivalent to death.In fact,if people can learn about tumors in advance and take active actions,nearly half of tumor cases can be avoided.In today’s medical knowledge informatization,although people can easily get what they need by searching,it contains a lot of inaccurate and even harmful information,which seriously affects people’s judgment.Therefore,the tumor knowledge recommendation system came into being.It can help the public to popularize the methods of tumor prevention,improve the quality of life of cancer patients,and at the same time help medical staff to acquire cutting-edge tumor knowledge and improve the level of treatment and care.At present,most of the mainstream algorithms of recommendation systems are based on collaborative filtering,that is,predictions are made by comparing the similarity of users or items.The accuracy is high but requires a certain data scale,and it is difficult to work for new users.To solve this problem,a hybrid prediction algorithm is proposed,and a tumor knowledge recommendation model is established,so as to better guide users to discover useful tumor knowledge.First,this paper predicts the classification of tumor knowledge by training the Attention model,and combines the user’s tumor disease codes and tumor knowledge labels to obtain a preliminary tumor knowledge recommendation set to implement recommendations for new users.Secondly,a collaborative filtering algorithm is introduced to predict the user’s tumor knowledge score from the two dimensions of the user’s registration information and historical reading records,so as to improve the recommendation effect.Finally,the prediction results of GRU-Attention and collaborative filtering are mixed and stored in the Redis cache for use in recommendation,thereby improving the response performance of the recommendation system.In recent years,the technology of small programs has begun to flourish.Compared with apps that need to be downloaded and installed before they can be used,small programs have the advantages of being ready to use and easy to develop.Among the many platforms that provide small program applications,WeChat carries a large number of users.Therefore,in order to make it easier for people to obtain tumor knowledge,this paper designs and builds a tumor knowledge recommendation system based on WeChat applet and SSM framework.It can achieve personalized tumor knowledge recommendation through data preprocessing and Word2 vec training word vectors,embedding Attention and collaborative filtering models,and solve people’s difficult problems in tumor knowledge retrieval.