Contino, Andrea (A.A. 2024/2025) Sentiment to strategy: leveraging forums discussions to guide automated trading decisions: an experimental study that analyzes whether reddit crowd sentiment can power profitable and automated long–short stock trading strategies. Tesi di Laurea in Advanced coding for data analytics, Luiss Guido Carli, relatore Alessio Martino, pp. 37. [Bachelor's Degree Thesis]

Preview

PDF (Full text) Download (599kB) | Preview

Abstract/Index

Reddit and WallStreetBets. End-to-end pipeline overview. High-level architecture and data flow. Key technologies and libraries. Assumptions, scope and ethical considerations. Reddit data acquisition & pre-processing. Subreddit scrape strategy and API configuration. Ticker detection and company-name mapping. Sentiment analysis with FinBERT. Sentiment analysis in financial text. FinBERT: domain-specific transformer. From raw probabilities to a single sentence score. Aggregating scores within a trading day. Casting to a date ticker matrix. Market data retrieval & portfolio construction. Design principles for daily trading. Hyperparameters selection. Transforming daily signals into tradeable weights. Prices, returns & equity curve. Empirical choice of TOP_N. Benchmarking the WallStreetBets strategy. Comparison selection. Building the random-trader baseline. Metrics and benchmark. Next step.

References

Bibliografia: pp. 34-36.

Thesis Type:	Bachelor's Degree Thesis
Institution:	Luiss Guido Carli
Degree Program:	Bachelor's Degree Programs > Bachelor's Degree Program in Management and Computer Science, English language (L-18)
Chair:	Advanced coding for data analytics
Thesis Supervisor:	Martino, Alessio
Academic Year:	2024/2025
Session:	Summer
Deposited by:	Alessandro Perfetti
Date Deposited:	13 Nov 2025 15:34
Last Modified:	13 Nov 2025 15:34
URI:	https://tesi.luiss.it/id/eprint/43843

Downloads

Repository Staff Only

View Item