MEANT: Multimodal Encoder for Antecedent Information
This work addresses stock market prediction for investors by providing a multimodal approach, but it is incremental as it builds on existing multimodal methods with a focus on temporal data.
The paper tackled the problem of attending to temporal data with multiple information types in stock market prediction by introducing the MEANT model and TempStock dataset, resulting in a performance improvement of over 15% on existing baselines.
The stock market provides a rich well of information that can be split across modalities, making it an ideal candidate for multimodal evaluation. Multimodal data plays an increasingly important role in the development of machine learning and has shown to positively impact performance. But information can do more than exist across modes -- it can exist across time. How should we attend to temporal data that consists of multiple information types? This work introduces (i) the MEANT model, a Multimodal Encoder for Antecedent information and (ii) a new dataset called TempStock, which consists of price, Tweets, and graphical data with over a million Tweets from all of the companies in the S&P 500 Index. We find that MEANT improves performance on existing baselines by over 15%, and that the textual information affects performance far more than the visual information on our time-dependent task from our ablation study.