Are you curious about machine learning and what it could mean to actuaries? The Modelling, Analytics and Insights from Data (MAID) working party has published a paper entitled Practical Application of Machine Learning Within Actuarial Work to bring machine learning to life.
Are you curious about machine learning and what it could mean to actuaries? The Modelling, Analytics and Insights from Data (MAID) working party has published a paper entitled 'Practical Application of Machine Learning Within Actuarial Work' to bring machine learning to life.
The paper provides an introductory overview of machine learning techniques, including some of the potential benefits of adopting them, and a few programming platforms to consider. Fundamental concepts on data processing (including non-traditional and unstructured data), model training and model validation are brought to life in four exploratory case studies. In doing so, parallels between the machine learning project cycle and the actuarial control cycle are introduced.
The paper concludes with a number of lessons learnt. Of specific relevance to actuaries is the importance of being able to apply these techniques effectively within an actuarial function by applying our expert domain knowledge. Machine learning techniques are essential toolkits in a world of high-volume and varied datasets. They are also more accessible than ever before via open source software. However, to fully use these techniques, a deep understanding of the business problem and constraints is vital.
The world of machine learning techniques is fast evolving, which means the work of the working party is far from complete. Phase two of the research has started to devise a practical guide on the end-to-end application of machine learning techniques, from defining/understanding the business problem to communicating and monitoring model performance. Watch this space.
Visit Modelling, Analytics and Insights from Data WP: bit.ly/2FxYEKz
