医保基金监管_个人骗保行为识别模型

通过医保结算数据，医保结算明细数据，门诊记录数据，医院住院数据，药店购药数据等，构造一定时间内个人购药次数、购药金额、购药次数同比增长率、住院次数，医保结算金额占比等特征，使用异常检测、聚类等算法，识别个人可能存在的使用医保基金骗保行为

This dataset leverages multi-source medical-related data including medical insurance settlement data, detailed medical insurance settlement records, outpatient medical records, hospital inpatient data and pharmacy medication purchase data to construct a series of features within a specified time period, such as an individual's number of medication purchases, total medication purchase amount, year-on-year growth rate of the number of medication purchases, number of hospitalizations, and proportion of medical insurance settlement amount. Subsequently, anomaly detection, clustering and other algorithms are applied to identify potential medical insurance fund fraud behaviors committed by individuals.

搜集汇总

背景与挑战

背景概述

该数据集整合医保结算、门诊住院及购药记录等多源数据，通过构建购药频次、金额变化等特征，采用异常检测和聚类算法识别个人骗保嫌疑。