Fraud Detection for Insurance Claims Processing () ;

Problem Solved

Traditional known Parameters used for Prediction

Manual Rules

Automating and Learning changing Behaviors

Solution

ML for Predictive Detection

Uncover Non-Intuitive Patterns and Subtle Trends

Automation for future Trends

Benefits

Reduce operational cost through Automation

Faster Claim Processing for Better Customer Experience

Better Risks Management through Identifying Customer Trends

Predict new kind of frauds with latest data

Approach

Tested on Large Data Set

PCA Transformation and Feature Scaling

Algorithm Used

  • Neural Networks
  • Random forest algorithm
  • Bayesian statistics
  • Support Vector Machine