INSURANCE CLAIM FRAUD DETECTION

Abstract

In this paper we are exploring the significance of Fraud claim Detection in insurance project 
itself is significance and novelty project because of the data set used for detecting fraud is very 
huge and imbalanced or skewed. And the different types of techniques used to balance the data 
set like re sampling methods for better result, finding of best algorithms to classify the dataset 
and detect fraud is a very challenging and the some of the issues faced using these techniques 
and methods. The imbalance in original data set and most of the claims being non fraud. If we 
use the original data set as the base for our predictive models and analysis there is a high chances 
of getting lots of errors and algorithms to over fit since it will assume that most the claims as non 
fraud. So choosing best techniques, methods and algorithms to get better result is a novelty in 
this project medical insurance fraud detection