StockMem: An Event-Reflection Memory Framework for Stock Forecasting
He Wang, Wenyilin Xiao, Songqiao Han, Hailiang Huang
TL;DR
StockMem tackles stock forecasting under noisy, real-time news by introducing a dual-layer memory that converts news into structured events and tracks their evolution to extract incremental signals. It combines an Event Memory for a temporally structured knowledge base with a Reflection Memory that captures causal experiences from event-price interactions, enabling analogical retrieval across sequences and across companies. Retrieval and inference are built on event-sequence similarity and cross-company references, delivering explainable predictions by tracing the information chain. Empirical results show StockMem outperforms strong baselines and ablations confirm the importance of structured events, longitudinal tracking, and cross-domain references for accurate, transparent forecasts.
Abstract
Stock price prediction is challenging due to market volatility and its sensitivity to real-time events. While large language models (LLMs) offer new avenues for text-based forecasting, their application in finance is hindered by noisy news data and the lack of explicit answers in text. General-purpose memory architectures struggle to identify the key drivers of price movements. To address this, we propose StockMem, an event-reflection dual-layer memory framework. It structures news into events and mines them along two dimensions: horizontal consolidation integrates daily events, while longitudinal tracking captures event evolution to extract incremental information reflecting market expectation discrepancies. This builds a temporal event knowledge base. By analyzing event-price dynamics, the framework further forms a reflection knowledge base of causal experiences. For prediction, it retrieves analogous historical scenarios and reasons with current events, incremental data, and past experiences. Experiments show StockMem outperforms existing memory architectures and provides superior, explainable reasoning by tracing the information chain affecting prices, enhancing decision transparency in financial forecasting.