Generalized Stock Price Prediction for Multiple Stocks Combined with News Fusion

TL;DR

This paper proposes a generalized stock price prediction model that combines large language models and news data, effectively predicting multiple stocks simultaneously with improved accuracy over traditional methods.

Contribution

It introduces a novel attention-based news filtering technique using stock name embeddings and trains a single model applicable to multiple stocks, unlike prior individual stock models.

Findings

Achieved a 7.11% reduction in MAE compared to baseline methods.

Utilized attention mechanisms for effective news relevance filtering.

Demonstrated the effectiveness of LLM-based news encoding in stock prediction.

Abstract

Predicting stock prices presents challenges in financial forecasting. While traditional approaches such as ARIMA and RNNs are prevalent, recent developments in Large Language Models (LLMs) offer alternative methodologies. This paper introduces an approach that integrates LLMs with daily financial news for stock price prediction. To address the challenge of processing news data and identifying relevant content, we utilize stock name embeddings within attention mechanisms. Specifically, we encode news articles using a pre-trained LLM and implement three attention-based pooling techniques -- self-attentive, cross-attentive, and position-aware self-attentive pooling -- to filter news based on stock relevance. The filtered news embeddings, combined with historical stock prices, serve as inputs to the prediction model. Unlike prior studies that focus on individual stocks, our method trains a single generalized model applicable across multiple stocks. Experimental results demonstrate a 7.11% reduction in Mean Absolute Error (MAE) compared to the baseline, indicating the utility of stock name embeddings for news filtering and price forecasting within a generalized framework.