Language Models for Financial News Recommendation

Abstract

We present a unique approach to identifying news stories that influence the behavior of financial markets. SpecificaUy, we describe the design and implementation of iEnalyst, a system that can recommend interesting news stories ~ stories that are likely to afl^ect market behavior. iEnalyst operates by correlating the content of news stories with trends in financial time series. We identify trends in time series using piecewise linear fitting and then assign labels to the trends according to an automated binning procedure. We use language models to represent patterns of language that are highly associated with particular labeled trends. iEnalyst can then identify and recommend news stories that are highly indicative of future trends. We evaluate the system in terms of its ability to recommend the stories that will afl^ect the behavior of the stock market. We demonstrate that stories recommended by iEnalyst could be used to profitably predict forthcoming trends in stock prices.