Big data science for predicting insurance claims fraud

Abstract

The cost of insurance is on the rise globally with contributing factors being new legislation and fraud. Both of these factors are problematic with new legislation not easily controllable by insurance companies and fraud very difficult to proactively detect. Hence, controlling the upward trend of cost is problematic. We therefore investigate the use of Big Data and Data Science (Big Data Science) to predict insurance claims fraud. Big Data, Data Science and Predictive Analytics were applied to the shortterm insurance industry with a use-case in the developing world. This study was performed within the bounds of the constraints of privacy legislation that invokes standards on how data can be kept and shared — we suggest a privacy preservation method of predicting insurance claims fraud.