India is a huge market for insurance but the industryr is bleeding losses due to increasing fraud from inside and outside of the system. Insurance frauds leads to close to INR 40,000 crores loss which is close to 8.5% revenue of the insurance sector in India. These huge losses have prompted insurance companies to set-up a new anti-fraud department whose job is to identify the risk and loss to these frauds and to nd out ways to reduce fraudulent claims. We now have a data from an insurance company (Name not disclosed) which is losing revenue in fraudulent claims. Based upon the given data build the below models: ' Predict if the DRV_CLA|M_STATUS will be 'Rejected' or 'Accepted'. Consider 'Closed' status given in data as 'Accepted' status. This prediction will help the Claims Management team to manage the claims effectively. Prediction model will give them the list of claims to be rejected in advance thus helping them to scrutinize these claims and ensure no leakage in for of frauds. 0 Identify clusters/associations in the claims data to classify them based on the severity of the claims. Identify elds that are classifying the claims as highly severe claims. These will be the claims that will need more focus and scrutiny from the fraud management team