SUBMIT
Request More Information

For more information, contact: 
Melini Hadjitheori
melinih@marcusevanscy.com

What is the outlook for machine-learning in personal line pricing? Where is the cutting edge?

Machine-learning is a very broad concept and spans supervised and unsupervised learning and a huge variety of mathematical and modelling approaches.  Personal Lines pricing has so far only scratched the surface in regression problems but already proven that we are in the middle of a paradigm change. Different decision tree methods outperform traditional GLM’s and as method evolve that allow us to understand and regulate models better this trend will accelerate even further. However, the most revolutionary changes are to be expected from neural nets and deep learning, where we already find new features and ways to model more effectively though gained insights into system properties.
 
How can insurance companies maximise the power of machine-based pricing models? 

Obviously, you have first to make sure that the basic ingredients are in place; a good supply of quality data and a powerful modelling environment. Our experience shows that you can then make a difference in two aspects. The first area is model sophistication. Independent of application you can improve prediction accuracy if you are prepared to test new methods. The ability to learn fast and put changes in place is the difference between companies that will be able to maximize the power of machine-based pricing and those who won’t. This is particularly important for fast-moving portfolios in personal lines since the losers do not only end up with less business but also with a worse risk selection. In my experience, you need to have quite sophisticated data science capabilities in-house and can use outside expertise in bleeding edge areas.
The other area is efficiency and we advocate a strategy that makes it easy to rerun, recalibrate and reestimate models quickly and automatically. Don’t expect machine-based pricing to produce new models without human interventions but use machine-based methods to keep your models up to date. This is a strategy that has to be built in a design phase in order to work smoothly in your particular environment.

What are the lessons learned with respect to the analysis of data?


Since the area of machine learning is rapidly evolving and most achievements are open-sourced, businesses must use an approach that is aligned with the rest of the community.
In our experience, you have to encourage a test and learn culture that keeps up to date with the latest academic research. This can be achieved through gamifying the analytical exploration of model challenges. Our main lesson learned is that it has to be fun to work analytically.

What would you like to achieve by attending the 2nd Annual Pricing in Personal Lines Insurance?

Many conferences have a very broad scope which makes it difficult to have a rewarding discussion that can be translated to tangible insights in the business. I expect that the 2nd Annual Pricing in Personal Lines insurance to be different and that I and my peers can exchange real experiences and learnings. Even if insights are often transferable and interesting in a broader sense and a variety of fields it is of value to be able to discuss them in the context of a particular application.

Ahead of the 2nd Annual Pricing in Personal Lines Insurance conference, we spoke with Dr. Sascha Firle, Director of Advanced Analytics at RSA-Scandinavia about how insurance companies maximise the power of machine-based pricing models.

To view the Conference Agenda, click HERE! 

About the Conference:

This marcus evans conference will look at how firms are enhancing personal lines pricing models with data analytics, machine based learning and AI to improve predictability and increase profitability. The 2nd Annual Pricing in Personal Lines Insurance conference will take place on the 17th to 19th September 2018 in
Marriott, West India Quay, Canary Wharf, London, United Kingdom.

To view the Conference Agenda, click HERE!

Copyright © 2018 Marcus Evans. All rights reserved.

About the speaker:

Dr Sascha Firle has a PhD in Theoretical Physics, Chaos Theory and wrote his thesis “Get a grip on chaos: Tailored measures for complex systems on surfaces in 1999. He has  20 years of experience from the P&C insurance business. Sascha has been the Head of Personal Line Pricing in one of the major Swedish insures for 3 years and is now Director of Advanced Analytics in RSA Scandinavia.

RISE OF THE MACHINES IN PERSONAL LINES PRICING 

An interview with the Director of Advanced Analytics in RSA Scandinavia

Dr Sascha Firle, Director Advanced Analytics, Nordic UW, Head of Modelling and Pricing
RSA

Speakers Include: 
  • Andrew Davies
    Head of Pricing Strategy, Delivery and Risk

    LV

  • Dan Bishop
    Head of Pricing
    Hiscox


  • Sascha Firle
    Director Advanced Analytics, Nordic UW, Head of Modelling and Pricing

    RSA

  • Steven Ball
    Head of Pricing and Analytics
    NFU Mutual


  • Martin Harding
    Head of IT Transformation
    Markerstudy


  • Julien Durand
    Senior Data-Scientist for Pricing
    AXA
Fix the following errors:
Hide