Motor Insurance is the largest contributor to insurance fraud in Kenya.In 2015 the Insurance Regulatory Authority reported that the insurance sector had lost KSh324.7 million to fraud. This project has attempted to develop a ML algorithm to detect.I have used the historical motor insurance claims data including normal motor insurance claims and fraud ones to obtain normal/fraud behavior features
based on machine learning techniques, and utilized these features to check if a motor insurance claims is fraud or not.I did a comparative study to decide which classifier is best for this project to train the behavior features of normal and abnormal motor insurance claims.
The objective is to construct a model to predict which motor insurance claims could be fraudulent with high accuracy.
The data that is extracted from one of the Insurance companies in Kenya.I created a predictive model that predicts if an insurance claim is fraudulent or not. The answer is between a YES/NO;a Binary Classification . I dealt with a classification algorithm: random forest model to detect fraudulent motor insurance claims in Python. The model is deployed with flask, therefore the model is an API to a website developed in laravel that facilitates users(clients) to register a claim that is then processed through the ML model to predict if its fraudulent or not.