Insurance data mining: algorithms, ethics, and security
The Canadian insurance company La Capitale Assurance et Services Financiers has accumulated over time several massive databases of information about their customers in their various insurance programs. They are now looking for a way to exploit this data to gain a better understanding of their customers, in order to customize their insurance coverage to each individual customer's needs. This must be done in the respect of strict ethical norms and within the constraints that the users themselves may impose on how the company can use their data. And, considering the sensitive and personal nature of the information obtained, data security must be extremely robust. This research program is thus aligned around three fundamental axes. The data mining axis aims to develop new algorithms to infer information about individual customers from data about their interactions with the company. The ethics and social acceptability axis studies customer expectations about this data mining, both regarding how they will benefit from it and how their private lives will be protected. And the data security axis develops new techniques to guarantee the confidentiality, integrity, availability, and traceability of the data. The insurance industry is a multi-billion dollar industry in the Canadian economy, and the protection of personal information is a topic of great importance for a majority of the Canadian population. The data management tools and knowledge discovery algorithms we will will give Canadian companies an important edge in this competitive international marketplace, while guaranteeing to the Canadian population control and protection of its personal information.
Co-applicants: Khoury, Lyse Langlois, Pierre-Luc Déziel
Collaborators: Nadia Tawbi, Josée Desharnais, Mohamed Mejri.