A Multilayer Perceptron based Ensemble Technique for Fine-grained Financial Sentiment Analysis

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 540–546 Copenhagen, Denmark, September 7–11, 2017. c⃝2017 Association for Computational Linguistics A Multilayer Perceptron based Ensemble Technique for Fine-grained Financial Sentiment Analysis Md Shad Akhtar, Abhishek Kumar, Deepanway Ghosal Asif Ekbal, Pushpak Bhattacharyya Department of Computer Science and Engineering Indian Institute of Technology Patna, India {shad.pcs15,abhishek.ee14,deepanway.me14,asif,pb}@iitp.ac.in Abstract In this paper, we propose a novel method for combining deep learning and classical feature based models using a Multi-Layer Perceptron (MLP) network for financial sentiment analysis. We develop various deep learning models based on Convolu- tional Neural Network (CNN), Long Short Term Memory (LSTM) and Gated Recur- rent Unit (GRU). These are trained on top of pre-trained, autoencoder-based, fi- nancial word embeddings and lexicon fea- tures. An ensemble is constructed by com- bining these deep learning models and a classical supervised model based on Sup- port Vector Regression (SVR). We eval- uate our proposed technique on a bench- mark dataset of SemEval-2017 shared task on financial sentiment analysis. The pro- pose model shows impressive results on two datasets, i.e. headlines datasets.microblogs and news Comparisons show that our proposed model performs better than the existing state-of-the-art systems for the above two datasets by 2.0 and 4.1 cosine points, respectively. 1 Introduction Microblog messages and news headlines are freely available on Internet in vast amount. Dynamic na- ture of these texts can be utilized effectively to analyze the shift in the stock prices of any com- pany (Goonatilake and Herath, 2007). By keep- ing a track of microblog messages and news head- lines for financial domain one can observe the trend in stock prices, which in turn, allows an in- dividual to predict the future stock prices. An First three student authors have equally contributed to this work increase in positive opinions towards a particular company would indicate that the company is do- ing well and this would be reflected in the increase in company stock prices and vice-versa. Benefits of such analysis are two-fold: (i). an individual can take informed decision before buying/selling his/her share; and (ii). an organization can utilize this information to forecast its economic situation. Sentiment prediction is a core component of an end-to-end stock market forecasting business model. Thus, an efficient sentiment analysis sys- tem is required for real-time analysis of financial text originating from the web. Sentiment analy- sis in financial domain offers more challenges (as compared to product reviews domains etc.) due to the presence of various financial and techni- cal terms along with numerous statistics. Coarse- level sentiment analysis in financial texts usually ignores critical information towards a particular company, therefore making it unreliable for the stock prediction. In fine-grained sentiment anal- ysis, we can emphasize on a given company with- out losing any critical information. For example, in the following tweet sentiment towards $APPL (APPLE Inc.) is positive while negative towards $FB (Facebook Inc.). ‘$APPL going strong; $FB not so.’ In literature, many methods for sentiment anal- ysis from financial news have …

Content truncated to 500 words