A Machine Learning Based Strategy for Election Result Prediction
A Machine Learning Based Strategy for Election Result Prediction
复制标题
DOI：
10.1109/csci49370.2019.00263
复制
发表时间：
2019-12
期刊：
影响因子：
--
通讯作者：
Meng-Hsiu Tsai;Yingfeng Wang;M. Kwak;N. Rigole
中科院分区：
文献类型：
--
作者：
Meng-Hsiu Tsai;Yingfeng Wang;M. Kwak;N. Rigole
预测选举结果是政治学中的一个热门领域。在过去十年中，社交媒体在政治选举中得到了广泛应用。大多数方法能够预测全国性选举的结果。然而，预测许多地方选举的总体结果仍然具有挑战性。本文提出了一种基于机器学习的策略，用于分析推特数据以预测许多地方选举的总体结果。为了验证这一策略的有效性，我们将其应用于分析基于2018年美国中期选举的推特数据。结果表明，预测结果与实际选举结果相近。
Predicting election results is a hot area in political science. In the last decade, social media has been widely used in political elections. Most approaches can predict the result of a national election. However, it is still challenging to predict the overall results of many local elections. This paper presents a machine learning based strategy to analyze Twitter data for predicting the overall results of many local elections. To verify the effectiveness of this strategy, we apply it for analyzing the Twitter data based on the 2018 midterm election in United States. The results suggest the predicted results are close to the actual election outcome.