Modelling dynamic policyholder behaviour through machine learning techniques

Modelling dynamic policyholder behaviour through machine learning techniques

305 Views
Added in AFIR / ERM ASTIN / NON-LIFE

Thanks! Share it with your friends!

URL

You disliked this video. Thanks for the feedback!

Sorry, only registred users can create playlists.
URL

Description:

Speaker(s): Marco Aleandri (University La Spienza)

In this study, we focus our attention on the most relevant non-market risk in insurance, that is, lapse risk. It is basically linked to the behavior of policyholders facing various situations such as aging, actual economic condition, contract features, and so on. At the same time, policyholder's retention directly impacts the profitability of the product itself, thus the profitability of the company as a whole. Through the first part of our analysis, we will recognize some relevant lapse risk factors from a specific dataset including a number of explanatory variables. More importantly, the predictive results from the traditional logistic regression will be compared to those of a bagging classification tree, in order to select the most powerful model. Furthermore, the goal of the second part of the analysis is the valuation of the impact on the profitability of a specific insurance product based on the predicted lapse rates. We will observe how significant the policyholder behavior can be as soon as it is introduced within the profit valuation in a dynamic fashion.

Post your comment

Sign in or sign up to post comments.
Be the first to comment