The client’s data was not labelled as fraudulent or not. Proposed a solution that would take up information that were filed during the claiming process and then predict whether the claim was fraudulent or not.
Initially the currently available mechanism, policies and regulations were studied.
Data is analyzed to examine and check for abnormal data points and comprehend variables that may cause anomalies.
Initially the duration between accident to the date of the report was analyzed.
Driver specific analysis was performed where in the license class, the experience and vehicle restriction and then visualized
The isolation Forest and SVM (Support Vector Machine) model was constructed to identify claims that stood apart as anomalous
Ability to identify claims that stood out as an anomaly.
Ability to identify the variables that cause abnormality