José Cristino Fernández, Business Director of Future Space, presented yesterday the paper "Graph Analysis for Fraud Detection" in the framework of the 3rd Annual Forum on Claims and Fraud Management organized by INESE, in which Future Space participated as a sponsor together with other companies.
Santiago Martín, Director of the CR Bulletin and Ángeles Romero, General Manager of BCN IURIS were in charge of moderating the conference in which the main focus was on new ways of handling claims in the insurance sector, taking into account the changes in clients, who are increasingly informed and demanding, as well as an economic upturn and the premise of being aware of the importance of managing claims and fraud.
In that sense, the following data are relevant:
- Last year there was an increase in fraud attempts on the most popular lines
- 2 out of 100 life insurance claims contain deception
- The auto industry still has the highest rate of fraud, 6 out of 100 injuries are illegitimate
- For every euro invested in fraud detection, 34.70 euros are recovered
As an innovative way of trying to solve this problem, José Cristino proposes to move from manual treatment to predictive models. In this way, new variables and unknown patterns are identified, bringing benefits such as increased productivity, enriched information and decreased occurrence of fraudulent claims. The specific solution to this approach is Graph Analysis.
SNA (Social Network Analysis) is a discipline of analysis whose objective is to be able to understand social structures through a network formed by nodes (individuals, telephones, addresses, accidents, etc.) and the relationships that exist between them. These analysis techniques are considered vital in disciplines such as sociology, biology, anthropology, among others, and can be used to understand any social structure.
This approach allows the detection of new patterns of behaviour in the company's historical information, with the aim of analysing these patterns for each new claim received by the company. And what does it bring to a company? Well, a visual environment based on graphs to improve the analysis process, extending the focus of the analysis of the claim to the whole of the company's claims and the identification of anomalies in the relationship patterns in the company's claims.
Any insurance company claim usually meets a certain standard. Grafos Analysis is capable of detecting these patterns automatically and continuously over time, therefore, the system is able to detect situations in which the normal parameters are not met.
As a conclusion, it is important to point out that, in a global way, Graph Analysis, being a highly visual tool, allows us to understand the relationships of our information, to explore the data from a new perspective, finding new clues. Similarly, the algorithms on graphs allow us to find patterns and anomalies of behavior in our information, providing new indicators.
Although there are no magic formulas and the imagination of fraudsters is infinite, with Analytics and the exploration of their information we have an invaluable ally in the fight against fraud, using technology and innovation as the way to bring value to each business.
