Short communication: Deep fundamental factor models

DOI10.1137/20M1330518MaRDI QIDQ5131409zbMATH OpenFDO

Authors Matthew F. Dixon, Nick Polson

Publication date 7 November 2020

Published in SIAM Journal on Financial Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1903.07677

zbMATH Keywords

deep learningfactor modelsfinancial econometricsinterpretability

Mathematics Subject Classification ID

Diagnostics, and linear inference and regression (62J20) Applications of statistics to actuarial sciences and financial mathematics (62P05) Artificial neural networks and deep learning (68T07) Statistical methods; risk measures (91G70)

Abstract: Deep fundamental factor models are developed to automatically capture non-linearity and interaction effects in factor modeling. Uncertainty quantification provides interpretability with interval estimation, ranking of factor importances and estimation of interaction effects. With no hidden layers we recover a linear factor model and for one or more hidden layers, uncertainty bands for the sensitivity to each input naturally arise from the network weights. Using 3290 assets in the Russell 1000 index over a period of December 1989 to January 2018, we assess a 49 factor model and generate information ratios that are approximately 1.5x greater than the OLS factor model. Furthermore, we compare our deep fundamental factor model with a quadratic LASSO model and demonstrate the superior performance and robustness to outliers. The Python source code and the data used for this study are provided.

Recommendations

• Autoencoder asset pricing models
• Neural network-based automatic factor construction
• Efficient factor GARCH models and factor-DCC models
• DYNAMIC FACTOR MODELS
• Deep learning for finance: deep portfolios

Cites work

• Deep learning for finance: deep portfolios
• Multilayer feedforward networks are universal approximators
• Sufficient forecasting using factor models

Cited in

(3)

• Autoencoder asset pricing models
• Neural network-based automatic factor construction
• Industrial Forecasting with Exponentially Smoothed Recurrent Neural Networks

Describes a project that uses

Uses Software

• GitHub

This page was built for publication: Short communication: Deep fundamental factor models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5131409)