Listening to Chaotic Whispers: A Deep Learning Framework for\n News-oriented Stock Trend Prediction

Abstract

Stock trend prediction plays a critical role in seeking maximized profit from\nstock investment. However, precise trend prediction is very difficult since the\nhighly volatile and non-stationary nature of stock market. Exploding\ninformation on Internet together with advancing development of natural language\nprocessing and text mining techniques have enable investors to unveil market\ntrends and volatility from online content. Unfortunately, the quality,\ntrustworthiness and comprehensiveness of online content related to stock market\nvaries drastically, and a large portion consists of the low-quality news,\ncomments, or even rumors. To address this challenge, we imitate the learning\nprocess of human beings facing such chaotic online news, driven by three\nprinciples: sequential content dependency, diverse influence, and effective and\nefficient learning. In this paper, to capture the first two principles, we\ndesigned a Hybrid Attention Networks to predict the stock trend based on the\nsequence of recent related news. Moreover, we apply the self-paced learning\nmechanism to imitate the third principle. Extensive experiments on real-world\nstock market data demonstrate the effectiveness of our approach.\n