Scoring and prediction of early retirement using machine learning techniques: application to private pension plans

Abstract

Artificial intelligence techniques have become very popular in public and private organizations since they allow a more accurate decision-making process. Private insurance companies have ventured into this field by implementing algorithms that allow a better understanding of available data. The knowledge of retirement decisions allows the insurance companies to detect early retirement at a given time so that they have the adequate budgetary provision in place. In this paper, machine learning algorithms and data from private pension plans are used to predict whether a person retires before or after 65 years old in function of both individual characteristics and macroeconomic factors.