Hybrid Neural-Classical Correction for Frozen Time Series Foundation Models: A Comprehensive Ablation Study on High-Frequency Stock Prediction

By Kasun Dewage, Suranadi De Silva, Shankhadeep Mondal

Published 2026-08-09

Everscope rating

1378.4

Relevance to quantitative trading

9 / 10

Implementation complexity

5 / 10

Reproducibility

5 / 5

About this paper

Methodology: Hybrid Neural-Classical Correction Framework. Problem types: Time Series Forecasting, Regression, Transfer Learning, Zero-shot Learning.

arXiv:2608.08825 · Code · Paper rankings

Open the interactive Everscope explorer for full analysis, charts, and paper battles.