The Design And Realization Of Recommending Item And Diagnosis In The Management System Of Physical Health Examination

In recent years, people began to get themselves into trouble because of the problem of information surplus. The situation that lack of information has gone. The recommendation information service that used widely changed from pull to push. This techonology make people get the information they want.Both the Bayesian network and the algorithm of K nearest neighbor are the main methods that used widely in recommendation system. Bayesian network is a recommendation mode that based probability. It has been used in many fields such as Diagnosis,prediction,military decisions,intelligent robots,Pathological diagnosis,financial market analysis, Data mining and so on. The algorithm of KNN is a nonparametric classification techonology, it is a mature method and effective in Pattern Recognition which is based on statistic.This thesis introduce the recommendation method into the 'my-verygood health examination management system' according to the problem that customers going to the physical examination center blind to choose the items and have to wait for a long time.First, construct the new customers' Bayes network recommendation modle which can recommend the items to the new customers in real-time. this method reduces the customers' blindness and time on choosing the items and solves the problem that waiting for a long time.Then,proposed two levels K nearest neighbor algorithm by improving the original K nearest neighbor.This method recommends diagnosis to the doctor in order to improve the diagnosis normative and decrease the input workload of doctor's,thus solve the problem of queueing.