Elementary Insurance

Elementary Insurance


Anomaly Detection for Fire Insurance

Significant increase in financial loss in comparison with increase of premium payments over the years draws attention. Accurate ratio is estimated as 1,3% of demands even though some abuses are detected.

Most common abuse types,

  • Prospect demands or demands from policies are detected in case of any damage or financial loss
  • Goal is to detect anomalies right before possible damages caused by weather conditions or other reasons.

It is possible to prevent abuses by detecting anomalies in fire insurance during production. Possible damages and anomalies are detected by using machine learning algorithms before they occur because of weather conditions or other reasons.

Main success points of preventing productions having high abuse risk

  • Preventing unnecessarily paid claims by detecting abuse
  • Creating profitable portfolio by working with a better agency portfolio
  • Getting to know portfolio of agencies and insurees better
Anomaly_Detection_for_Fire_Insurance-JFORCE