Cut the Chit-Chat: A New Framework for the Application of Generative Language Models for Portfolio Construction

28 Pages Posted: 31 Dec 2024

See all articles by Francesco Fabozzi

Francesco Fabozzi

Yale School of Management's International Center for Finance

Ionut Florescu

Stevens Institute of Technology - School of Business

Date Written: October 28, 2024

Abstract

Current applications of generative language models (GLMs) for portfolio construction utilize the chat functionality of such models to forecast expected returns. The current literature classify outputs of these models using discrete labels that ignore the magnitude of sentiment. We show that this procedure is not optimal for cross-sectional portfolio construction. In this paper, we introduce Logit Extraction, a methodology for extracting the model's assigned probabilities to sentiment labels, allowing for the formulation of a continuous-valued ranking variable for portfolio construction. We demonstrate that Logit Extraction significantly enhances risk-adjusted returns over previous discrete label approaches. We make available an open source implementation of Logit Extraction in the python package "TokenProbs".

Keywords: large language models, news sentiment, portfolio construction, generative AI

JEL Classification: G11, C45, G17

Suggested Citation

Fabozzi, Francesco and Florescu, Ionut, Cut the Chit-Chat: A New Framework for the Application of Generative Language Models for Portfolio Construction (October 28, 2024). Available at SSRN: https://ssrn.com/abstract=5002118 or http://dx.doi.org/10.2139/ssrn.5002118

Francesco Fabozzi (Contact Author)

Yale School of Management's International Center for Finance ( email )

165 Whitney Ave
New Haven, CT 06511
United States

Ionut Florescu

Stevens Institute of Technology - School of Business ( email )

Hoboken, NJ 07030
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
467
Abstract Views
1,153
Rank
127,125
PlumX Metrics