@InProceedings{balashankar-chakraborty-subramanian:2018:W18-31,
  author    = {Balashankar, Ananth  and  Chakraborty, Sunandan  and  Subramanian, Lakshminarayanan},
  title     = {Unsupervised Word Influencer Networks from News Streams},
  booktitle = {Proceedings of the First Workshop on Economics and Natural Language Processing},
  month     = {July},
  year      = {2018},
  address   = {Melbourne, Australia},
  publisher = {Association for Computational Linguistics},
  pages     = {62--68},
  abstract  = {In this paper, we propose a new unsuper- vised learning framework to use news events for predicting trends in stock prices. We present Word Influencer Networks (WIN), a graph framework to extract longitudinal tem- poral relationships between any pair of infor- mative words from news streams. Using the temporal occurrence of words, WIN measures how the appearance of one word in a news stream influences the emergence of another set of words in the future. The latent word-word influencer relationships in WIN are the build- ing blocks for causal reasoning and predic- tive modeling. We demonstrate the efficacy of WIN by using it for unsupervised extraction of latent features for stock price prediction and obtain 2 orders lower prediction error com- pared to a similar causal graph based method. WIN discovered influencer links from seem- ingly unrelated words from topics like poli- tics to finance. WIN also validated 67% of the causal evidence found manually in the text through a direct edge and the rest 33% through a path of length 2.},
  url       = {http://www.aclweb.org/anthology/W18-3109}
}