This paper focuses on the use of Big Data Analytics to detect motor insurance fraud.
Fraud Management is inevitable to insurance companies as this informs the level of risk covered which in turn affects the premiums being charged. Over the years, the effect of fraud on the insurance industry has increased immensely. Fraudsters are continuously perfecting their technique and applying emerging technologies to perfect their crime. Companies can no longer rely on their policyholders to uphold the utmost good faith principle or individual scrutiny of claims once reported, as they used to before. Millions have already been lost and measures have to be put in place to avoid the loss of even more.
Big data helps insurers have a superior understanding of their entire business environment and their client behaviours in all dimensions which informs fraud management techniques. This is made possible by the use of big data analytics tools. This paper uses machine learning; support vector machines (SVM) to classify claims as fraudulent or not. The model is also used to identify the claim characteristics most relevant in identifying fraudulent claims. The data used is motor insurance claims data consists of 1000 policyholders and 37 parameters.
The results showed that policyholder details (occupation and hobbies), accident details (incident state and number of vehicles involved), are the most relevant variables in fraud detection. Consistent with past research age and premium amounts were not relevant in classifying the claims as fraudulent. Comparing SVM and random forest methodologies, SVM was noted to be more accurate hence insurance companies can implement them in their fraud detection platforms.