Alex Marcuson presented on machine learning at the annual GIRO Conference, this year held during September in Dublin. Speaking together with Alan Chalk, the popular workshop provided an introduction for actuarial newcomers to this field.

The workshop provided a practical example of decision tree tools to analyse sample text. Using speeches given by US presidential candidates Hillary Clinton and Donald Trump, we demonstrated the accessibility of the techniques to actuaries. By explaining the underlying modelling process used, we hoped that attendees would be encouraged to try out the techniques at their workplaces.

A key message from the talk was the need for actuaries to engage with and provide leadership in the machine learning field. Reframing existing actuarial tools into a machine learning language would enable actuaries to be ready for a potentially rapid field of change in the insurance industry, rather than be rendered obsolete by it.

To conclude their talk, Alex and Alan held a Q&A session looking at some of the thorny professional and ethical questions thrown up by emerging techniques, and anticipating some of the challenges ahead.

GIRO 2016 Workshop B2 Slides