News Sentiment Embeddings for Stock Price Forecasting

Ayaan Qayyum · 2025

Read News Sentiment Embeddings for Stock Price Forecasting by Ayaan Qayyum (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-01970.txt.

Description

Can the tone of news heads foretell where a stock fund goes next? Ayaan Qayyum tests that idea in a 2025 study of a fund that tracks the five hundred largest listed firms in the United States. Headlines come from a major financial paper, and each head is turned into a set of numbers by a text model from OpenAI.

A squeeze step pulls the key traits from those numbers. Extra data on the dollar and on bond yields join the mix to lift accuracy. More than 390 model variants were trained in all. Early results say the head-based numbers boost forecasts by at least forty percent against models without them. The hard parts are lag effects, noise, and the fine shades of how news moves prices over days.

Catalog record

CREATORAyaan Qayyum
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-01970

Browse the full library · About the Tower of Babel