FinTSBridge: A New Evaluation Suite for Real-world Financial Prediction with Advanced Time Series Models

Yanlong Wang, Jian Xu, Tiantian Gao, Hongkang Zhang, Shao-Lun Huang, Danny Dongning Sun, Xiao-Ping Zhang

Keywords: Time Series, Financial Dataset, Benchmark

Abstract: Despite the growing attention to time series forecasting in recent years, many studies have proposed various solutions to address the challenges encountered in time series prediction, aiming to improve forecasting performance. However, the strategic interactions inherent in financial markets make it challenging to effectively apply time series forecasting models to asset pricing. There is still a need for a bridge to connect cutting-edge time series forecasting models with financial asset pricing. To bridge this gap, we have undertaken the following efforts: 1) We constructed three datasets from the financial domain; 2) We selected over ten time series forecasting models from recent studies and validated their performance in financial time series; 3) We developed new metrics, msIC and msIR, to quantify specific aspects of time series correlation captured by the models; 4) Task designs tailored to our financial data collections enabled rigorous evaluation of models' operational effectiveness in financial decision-making contexts. Our findings suggest that time series forecasting models, through domain-specific processing and evaluation, exhibit effective performance across diverse asset classes and sampling frequencies. We hope the developed new evaluation suite, FinTSBridge, can provide valuable insights into the effectiveness and robustness of advanced forecasting models for financial time series.

Dataset URL: https://www.kaggle.com/datasets/timalex/fintsbridge

Code URL: https://anonymous.4open.science/r/FinTSBridge-0D6C

Primary Area: Social and economic aspects of datasets and benchmarks in machine learning (e.g., fairness, interpretability, human-AI interaction, privacy, safety, strategic behavior)

Submission Number: 2468