Beyond theoretical data science: application to actuarial work: lapse analysis

Thursday 12th March 2020 12 PM GMT

How can life insurers address low persistency? How can data and analytics help?

We will demonstrate, with the use of a practical case study how the full cycle of actuarial analysis is evolving - from data collection and data enhancement, feature engineering, modelling, verification and ultimately application and communication.

We will provide an example application of data science applied to actuarial work including:

• Data preparation

• Visualisation

• Model fitting

• Validation

• Interpretation of results

• Performance of fitted data science model relative to one or two alternative models.

The case study provides an introduction to generalised linear modelling and its use as a predictive

modelling technique within insurance. We used open source programming software.

We will also touch on the potential to use external data, together with insurance companies data to

build an enhanced persistency model of their book of contracts.

‍

‍