Leveraging Machine Learning in Portfolio Construction

Exponential improvements in computer power and wide-spread availability of very large datasets has meant that AI and Machine Learning techniques are increasingly being applied to various problems in finance.

One important application is in portfolio construction. The FT Wilshire NxtGen Index series aims t solve the classic portfolio optimization problem first set out by Markowitz in the 1950s. The key advance has been achieved by leveraging Machine Learning to yield better estimates of predicted return.

In this paper we introduce various performance and implementation properties of the FT Wilshire NxtGen Index series. We also provide some insights into how performance improvements arise