PT  - JOURNAL ARTICLE
AU  - Derek Snow
TI  - Machine Learning in Asset Management: <em>Part 2: Portfolio Construction—Weight Optimization</em>
AID  - 10.3905/jfds.2020.1.029
DP  - 2020 Mar 12
TA  - The Journal of Financial Data Science
PG  - jfds.2020.1.029
4099  - https://pm-research.com/content/early/2020/03/12/jfds.2020.1.029.short
4100  - https://pm-research.com/content/early/2020/03/12/jfds.2020.1.029.full
AB  - This is the second in a series of articles dealing with machine learning in asset management. This article focuses on portfolio weighting using machine learning. Following from the previous article (Snow 2020), which looked at trading strategies, this article identifies different weight optimization methods for supervised, unsupervised, and reinforcement learning frameworks. In total, seven submethods are summarized with the code made available for further exploration.TOPICS: Big data/machine learning, analysis of individual factors/risk premia, portfolio construction, performance measurementKey Findings• Machine learning can help with most portfolio construction tasks, such as idea generation, alpha factor design, asset allocation, weight optimization, position sizing, and the testing of strategies.• This is the second of a series of articles dealing with machine learning in asset management and more narrowly weight optimization strategies equipped with machine learning.• Following from the previous article, different weight optimization methods are considered for supervised, unsupervised, and reinforcement learning frameworks.