LLM-Powered Sentiment Analysis for Stock Price Prediction

Notani, Mohit (2025) LLM-Powered Sentiment Analysis for Stock Price Prediction. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (2MB) | Preview

Abstract

The paper is an initial analysis of the sentiment analysis based on the FinBERT coupled with the Facebook Prophet time-series data involving enhancement of stock prediction. The experiment uses the additive modelling framework of Prophet complemented with sentiment regressors to make a one-day ahead forecast of Apple and Google prices as well as Tesla prices over three-years (2022-2025).The model is simply Prophet as the forecasting engine, with technical indicators used as external regressors on all stocks and additional sentiment features gained by using FinBERT on Tesla. There are 19,114 data about news headlines as input to the FinBERT sentiment analysis and the daily OHLCV market data. To optimize hyperparameters, it only runs over Prophets change point prior scale and seasonality prior scale parameters. Evaluation with 30-day out-of-sample testing shows that Prophet with sentiment reduces RMSE by 39.2 percent (on Tesla, the improvement is 14.38-8.74) and R2 by 1.85-fold (0.745- 0.311) against baseline Prophet models that use only technical features. Outcomes confirm the ability of the extra regressor framework proposed by Prophet to embed external text tuition simultaneously with interpretability of the underlying model that is crucial in financial use.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Horn, Christian UNSPECIFIED
Subjects:	H Social Sciences > HG Finance Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HG Finance > Investment > Stock Exchange
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	02 Jul 2026 14:11
Last Modified:	02 Jul 2026 14:11
URI:	https://norma.ncirl.ie/id/eprint/9442

Actions (login required)

View Item