Jean Dessain
Adv. Artif. Intell. Mach. Learn., 3 (4):1640-1653
Jean Dessain : IESEG School of Management Department of Finance 3 rue de la Digue 59000 Lille, France
DOI: https://dx.doi.org/10.54364/AAIML.2023.1193
Article History: Received on: 02-Sep-23, Accepted on: 14-Nov-23, Published on: 21-Nov-23
Corresponding Author: Jean Dessain
Email: j.dessain@ieseg.fr
Citation: Jean Dessain (2023). Improving the prediction of asset returns with machine learning by using a custom loss function. Adv. Artif. Intell. Mach. Learn., 3 (4 ):1640-1653
Not all errors from models predicting asset returns are equal in terms of impact on the
efficiency of the algorithm: a small error could trigger poor investment decisions while
a significant error has no financial consequences. This economic asymmetry, critical for
assessing the performance of algorithms, can usefully be replicated within the machine
learning algorithms itself through the loss function to improve its prediction capability. .
In this article: (a) we analyze symmetric and asymmetric loss functions for deep learning
algorithms. We develop custom loss functions that mimic the asymmetry in economic
consequences of prediction errors. (b) We compare the efficiency of these custom loss
functions with MSE and the linear-exponential loss “LinEx”. (c) We present an efficient
custom loss function that significantly improves the prediction of asset returns with improved
risk-return metrics (like Sharpe ratio twice better), and which we confirm to be
robust.