Updated Karen Clark model lets users “mix and match”

The latest catastrophe model release from Karen Clark & Co will allow users to “mix and match” components to customise their own model building process.

According to the company, its RiskInsight Version 4.4 (RI4.4) model includes enhancements to the custom model building capabilities and client integration modules, and performance optimisation.

It says that RiskInsight Version 4.4 advances the practice of building high-resolution custom catastrophe models with a transparent and modular design. RI4.4 makes it easier to import custom events and event intensity files and to visualise and verify the hazard data. RI4.4 also provides “plug-ins” for creating damage functions based on attributes that are not typically utilised in the traditional models.

The new model building capabilities make it easier to “mix and match” model components. The company says it will allows users to more efficiently customise the hurricane model damage functions and to use these customised functions with the Reference Model event catalogue.

According to Karen Clark, Hurricanes Hermine and Matthew demonstrated the live event capabilities of RiskInsight and how insurers can leverage these unique features to obtain reliable real time loss information. Forecast track information is updated every three hours, and in the latest update the live event tracking capabilities have been further enhanced.

“Significant investment has gone into making RiskInsight a highly advanced model building platform,” said Dr. Nozar Kishi, Karen Clark & Co Vice President, Model Development.

“In other modelling companies, the client software applications are separate from the model development activities so a lot of time is spent getting the peril models onto the software platforms. Karen Clark & Co’s leading scientists and engineers can build the model components directly on the RiskInsight platform, making the process highly efficient, and perhaps more important, the powerful visualisation tools enable detailed model verification and validation at every step in the process.”