Stock Index Prediction with Multi-task Learning and Word Polarity Over Time

TL;DR

This paper introduces a multi-task learning approach using BERT to predict stock index movements by extracting market sentiment and word polarity over time, addressing issues of relevance and polarity shifts in financial news analysis.

Contribution

It proposes a novel two-stage system with a sentiment extractor and summarizer, utilizing BERT with multitask learning and a new Polarity-Over-Time metric for improved stock prediction.

Findings

Effective prediction of weekly stock index movement.

Demonstrates the usefulness of Polarity-Over-Time metric.

Introduces a new 10-year Reuters financial news dataset.

Abstract

Sentiment-based stock prediction systems aim to explore sentiment or event signals from online corpora and attempt to relate the signals to stock price variations. Both the feature-based and neural-networks-based approaches have delivered promising results. However, the frequently minor fluctuations of the stock prices restrict learning the sentiment of text from price patterns, and learning market sentiment from text can be biased if the text is irrelevant to the underlying market. In addition, when using discrete word features, the polarity of a certain term can change over time according to different events. To address these issues, we propose a two-stage system that consists of a sentiment extractor to extract the opinion on the market trend and a summarizer that predicts the direction of the index movement of following week given the opinions of the news over the current week. We adopt BERT with multitask learning which additionally predicts the worthiness of the news and propose a metric called Polarity-Over-Time to extract the word polarity among different event periods. A Weekly-Monday prediction framework and a new dataset, the 10-year Reuters financial news dataset, are also proposed.