Enhancing Financial Market Prediction through Structured Prompt Engineering: A Comparative Analysis of Zero-Shot, Few-Shot, and Chain-of-Thought Sentiment Analysis

Venkataraman, Aashritha (2025) Enhancing Financial Market Prediction through Structured Prompt Engineering: A Comparative Analysis of Zero-Shot, Few-Shot, and Chain-of-Thought Sentiment Analysis. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (807kB) | Preview

Abstract

The aim of the study is to investigate the integration of better sentiment analysis techniques that entail systematic prompt engineering techniques with the view of improving the models that predict the financial market. The systematic method of the paper is to compare the techniques of the zero-shot, few-shot and chain-of-thought prompting techniques with the sentiment analysis, which could be conducted according to the text analysis of the DJIA market data that includes 300 trading days. With the XGBoost, FinBERT, ChatGPT ensemble methods, and multifaceted feature significance examination, the mission demonstrates that prompt engineering is quite useful as per outcomes of its being adequately performed in an orderly way. The ChatGPT ensemble resulted in an accuracy rate of 63.33 percent against the 55 percent baseline accuracy providing 8.33 percent of increments in market direction predictions and individual analysis of FinBERT seeking few-shot resulted in an accuracy of 60.0 percent with all three sentiments features receiving the top-10 rankings of feature importance. Observably, chain-of-thought sentiment is the third-most important feature (0.0674) which vividly illustrates that the chains of thought prompted yet highly valuable individual features of prediction. The findings supplement recent studies in the financial prompt engineering area by being the first investigation to compare techniques of prompting between ensemble and individual model frameworks in a systematic manner.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Chikkankod, Arjun UNSPECIFIED
Uncontrolled Keywords:	Sentiment Analysis; Prompt Engineering; Financial Prediction; Natural Language Processing; Machine Learning; Market Direction Prediction
Subjects:	H Social Sciences > HG Finance Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	03 Jul 2026 11:28
Last Modified:	03 Jul 2026 11:28
URI:	https://norma.ncirl.ie/id/eprint/9466

Actions (login required)

View Item