A Hybrid Approach to fraud detection on Health Insurance Claims.

      Аннотация: Data Mining techniques are holding out a great promise as regards their ability to improve detection of fraud and abuse. Data Mining combines powerful analytical techniques such as the improved K-means clustering and Multilayer Perceptron with knowledge to turn the data already acquired into the information and insight needed to identify probable instances of fraud. The textbook consists of five chapters and they are organized as follows: Chapter one, presents the introduction. The literature review on data mining is presented in chapter two with a comprehensive comparison between descriptive and prescriptive data mining on fraud detection. Chapter three, describes the methodology used in identifying the extent of fraud in health insurance industry and how the data was collected and preprocess. The software requirements are presented in chapter four. The results of the generic dataset and empirical datasets used are also presented and discussed. In chapter five, the conclusion and the recommendations are presented. The study established an improved Real-time assignment K-means clustering and multilayer perceptron hybrid approach to solve fraud detection problems.

Ключевые слова: artificial neural network, Data Mining, fraud detection, K-means clustering, Multilayer perceptron, Health Insurance claims