Loading...

Election vote share prediction using a sentiment-based fusion of Twitter data with Google trends and online polls

Kassraie, P ; Sharif University of Technology | 2017

623 Viewed
  1. Type of Document: Article
  2. Publisher: SciTePress , 2017
  3. Abstract:
  4. It is common to use online social content for analyzing political events. Twitter-based data by itself is not necessarily a representative sample of the society due to non-uniform participation. This fact should be noticed when predicting real-world events from social media trends. Moreover, each tweet may bare a positive or negative sentiment towards the subject, which needs to be taken into account. By gathering a large dataset of more than 370,000 tweets on 2016 US Elections and carefully validating the resulting key trends against Google Trends, a legitimate dataset is created. A Gaussian process regression model is used to predict the election outcome; we bring in the novel idea of estimating candidates' vote shares instead of directly anticipating the winner of the election, as practiced in other approaches. Applying this method to the US 2016 Elections resulted in predicting Clinton's majority in the popular vote at the beginning of the elections week with 1% error. The high variance in Trump supporters' behavior reported elsewhere is reflected in the higher error rate of his vote share. © Copyright 2017 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved
  5. Keywords:
  6. Election prediction ; Gaussian process regression ; Google trends ; Sentiment analysis ; Social media text mining ; Twitter ; Data mining ; Forecasting ; Gaussian distribution ; Gaussian noise (electronic) ; Natural language processing systems ; Regression analysis ; Gaussian process regression ; Google trends ; Sentiment analysis ; Text mining ; Twitter ; Social networking (online)
  7. Source: 6th International Conference on Data Science, Technology and Applications, DATA 2017, 24 July 2017 through 26 July 2017 ; 2017 , Pages 363-370 ; 9789897582554 (ISBN)
  8. URL: https://www.scitepress.org/Papers/2017/64843/pdf/index.html