A bed-hanging behavior monitoring method based on characterizing patient characteristics

Abstract

The invention discloses a method for characterizing patients' characteristics and a monitoring method for bed-hanging behavior based on the method for solving the problem of fraudulent insurance behavior in medical insurance. The combination of statistics and machine learning can effectively describe the patient's hospitalization behavior and reduce the workload of manual identification. Characterization of patients is carried out from the characteristics of hospitalization and basic personal information of patients. The bed-hanging behavior detection method builds a bed-hanging behavior model, which is constructed by four processes of feature description-cluster analysis-manual screening-machine learning model building. The bed hanging behavior detection method uses the combination of statistics and machine learning to effectively identify the bed hanging behavior, characterize the patient's hospitalization behavior, and effectively characterize the patient. Using the clustering method, outliers are screened out and clustered It can greatly reduce the normal data and reduce the workload in the manual screening stage; after the bed-hanging behavior model is established, use this model to predict the patient's hospitalization behavior.

Description

technical field [0001] The invention relates to a method for characterizing patient characteristics and a method for detecting bed-hanging behavior based on the method for characterizing patients. Background technique [0002] With the development of society and economy, in order to provide a better medical insurance environment for the people, the country has made better and better medical insurance policies. However, there are always some people who defraud medical insurance through some false medical treatment behaviors, seriously damaging the rights and interests of other people who really need medical insurance treatment. Generally, medical insurance personnel's medical data can be used to help social security institutions to further detect abnormal data, but hospitals generate a large amount of original medical insurance data every day. If abnormal detection is only done manually, it will not only be slow and waste manpower, but also difficult to guarantee. Accuracy. ...

AI Technical Summary

Problems solved by technology

In many cases, it is not known in advance how many categories a given data set should be divided into. The selection of the initial clustering center has a great impact on the clustering results. Once the initial value is not well selected, it may not be possible to obtain effective clustering. Clustering results; the K-means algorithm needs to continuously adjust the sample classification and continuously calculate the adjusted new cluster center, so when the amount of data is very large, the time overhead of the algorithm is very large

Images