Ontologie jako narzędzie budowy modeli w ubezpieczeniowych systemach informacyjnych – ekstrakcja wiedzy ubezpieczeniowej ze źródeł internetowych
Ontologies as a tool for constructing models in insurance information systems – extraction of insurance knowledge from internet sources

Summary/Abstract: Extraction of insurance pricing models from internet sources consists in constructing a representation of the source and characteristic features of the extracted model used to determine re- lationships between values of variables which are independent of the model, and result from the above-mentioned characteristic features, and the insurance premium amount. Our aim is to obtain a model with a possibly small error calculated as a difference between predicted and actual values. At the same time, we optimise the model reconstruction process which consists in minimising the number of questions sent to the source. Using the information extraction, we actually do not just make use of the database made available by the website, but rather try to identify, on the basis of the data collected, the algorithm (parameters and function form) according to which the pre- mium amount is calculated. Therefore, in contrast to extraction of information from deep internet sources, in the approach proposed we deal with extraction of knowledge. In the part presented, we discuss in detail the method for reconstructing the model, the challenges it involves, and demonstrate a tool which supports its application. Furthermore, we present the results obtained and discuss the scope of application.

• Details
• Contents