MarketGPT: Developing a Pre-trained transformer (GPT) for Modeling Financial Time Series

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wheeler, Aaron, Varner, Jeffrey D.
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909403944321024
author Wheeler, Aaron
Varner, Jeffrey D.
author_facet Wheeler, Aaron
Varner, Jeffrey D.
contents This work presents a generative pre-trained transformer (GPT) designed for modeling financial time series. The GPT functions as an order generation engine within a discrete event simulator, enabling realistic replication of limit order book dynamics. Our model leverages recent advancements in large language models to produce long sequences of order messages in a steaming manner. Our results demonstrate that the model successfully reproduces key features of order flow data, even when the initial order flow prompt is no longer present within the model's context window. Moreover, evaluations reveal that the model captures several statistical properties, or 'stylized facts', characteristic of real financial markets and broader macro-scale data distributions. Collectively, this work marks a significant step toward creating high-fidelity, interactive market simulations.
format Preprint
id arxiv_https___arxiv_org_abs_2411_16585
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MarketGPT: Developing a Pre-trained transformer (GPT) for Modeling Financial Time Series
Wheeler, Aaron
Varner, Jeffrey D.
Trading and Market Microstructure
This work presents a generative pre-trained transformer (GPT) designed for modeling financial time series. The GPT functions as an order generation engine within a discrete event simulator, enabling realistic replication of limit order book dynamics. Our model leverages recent advancements in large language models to produce long sequences of order messages in a steaming manner. Our results demonstrate that the model successfully reproduces key features of order flow data, even when the initial order flow prompt is no longer present within the model's context window. Moreover, evaluations reveal that the model captures several statistical properties, or 'stylized facts', characteristic of real financial markets and broader macro-scale data distributions. Collectively, this work marks a significant step toward creating high-fidelity, interactive market simulations.
title MarketGPT: Developing a Pre-trained transformer (GPT) for Modeling Financial Time Series
topic Trading and Market Microstructure
url https://arxiv.org/abs/2411.16585