Exploiting Social Media for Stock Market Prediction with Factorization Machine

Abstract

When the stock market has become more and more competitive, the stock market prediction has been a hot research topic. Traditional methods are based on historical stock data, which ignore the latest market information. Later although financial news is proposed to access market information, there are some disadvantages for news to predict the stock market. Recently when the micro-blogging service has grown to a popular social media and provides a number of real-time messages for a lot of users, social media is proposed for the stock market prediction. In that case, the high-dimension of textual feature poses a main challenge. In this study, we propose a novel kind of model, Factorization Machine (FM), to predict the trend of stock market. FM not only alleviates the impact of high dimensionality, but also captures some aspects of basic linguistics. Experiments on real-world data show that FM can achieve 81% accuracy and get significantly more profit than state-of-the-art models. In addition, we shed light on how textual representation influences the prediction and find that FM is stable, which is applicable to other social media or prediction application generally.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 33 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.