Financial Fine-tuning a Large Time Series Model

TL;DR

This paper explores fine-tuning a large time series model, TimesFM, on financial data to improve market price predictions and trading performance, demonstrating its effectiveness over benchmarks.

Contribution

It introduces a financial fine-tuning approach for TimesFM, enhancing its accuracy and trading outcomes on diverse financial instruments.

Findings

Fine-tuned TimesFM outperforms baseline in price prediction accuracy.

The model achieves better trading returns and risk metrics.

Financial fine-tuning improves model robustness in market prediction.

Abstract

Large models have shown unprecedented capabilities in natural language processing, image generation, and most recently, time series forecasting. This leads us to ask the question: treating market prices as a time series, can large models be used to predict the market? In this paper, we answer this by evaluating the performance of the latest time series foundation model TimesFM on price prediction. We find that due to the irregular nature of price data, directly applying TimesFM gives unsatisfactory results and propose to fine-tune TimeFM on financial data for the task of price prediction. This is done by continual pre-training of the latest time series foundation model TimesFM on price data containing 100 million time points, spanning a range of financial instruments spanning hourly and daily granularities. The fine-tuned model demonstrates higher price prediction accuracy than the baseline model. We conduct mock trading for our model in various financial markets and show that it outperforms various benchmarks in terms of returns, sharpe ratio, max drawdown and trading cost.