SentiHeros at SemEval-2017 Task 5: An application of Sentiment Analysis on Financial Tweets

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 857–860, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics ACL 2017 Submission 116. Confidential review Copy. DO NOT DISTRIBUTE. 000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050 051 052 053 054 055 056 057 058 059 060 061 062 063 064 065 066 067 068 069 070 071 072 073 074 075 076 077 078 079 080 081 082 083 084 085 086 087 088 089 090 091 092 093 094 095 096 097 098 099 SentiHeros at SemEval-2017 Task 5: An application of Sentiment Analysis on Financial Tweets Narges Tabari, Armin Seyeditabari, Wlodek Zadrozny Narges Tabari: nseyedit@uncc.edu Armin Seyeditabari: sseyedi1@uncc.edu Wlodek Zadrozny: wzadrozn@uncc.edu Abstract Sentiment analysis is the process of identi- fying the opinion expressed in text. Re- cently it has been used to study behavioral finance, and in particular the effect of opinions and emotions on economic or fi- nancial decisions. SemEval-2017 task 5 focuses on the financial market as the do- main for sentiment analysis of text; specif- ically, task 5, subtask 1 focuses on finan- cial tweets about stock symbols. In this paper, we describe a machine learning classifier for binary classification of finan- cial tweets. We used natural language pro- cessing techniques and the random forest algorithm to train our model, and tuned it for the training dataset of Task 5, subtask 1. Our system achieves the 7th rank on the leaderboard of the task. 1 Introduction The recent explosion of textual data creates an unprecedented opportunity for investigating peo- ple’s emotions and opinions, and for understand- ing human behavior. Although there are several methods to do this, sentiment analysis is an espe- cially effective method of text categorization that assigns emotions to text (positive, negative, neu- tral, etc.). Sentiment analysis methods have been used widely on blogs, news, documents and mi- croblogging platforms such as Twitter. Although social media and blogging are pop- ular and widely used platforms to discuss many different topics, they are challenging to analyze. This is to large extent due to the specific of vo- cabulary and syntax, which are dependent on top- ics, with the same words possibly expressing dif- ferent sentiments in different contexts. For exam- ple, a word in a casual context might have positive or neutral sentiment (e.g., crush), while the same word generally has a negative sentiment in fi- nance. Therefore, with the absence of general nat- ural language understanding, context-dependent and domain-specific approaches allow us to in- crease the accuracy of sentiment analysis at a rela- tively low implementation cost. Domain-specific sentiment analysis is being used to analyze or investigate various areas in fi- nance, such as corporate finance and financial markets, investment and banking, asset …

Content truncated to 500 words