Analisis Sentimen Persepsi Masyarakat Terhadap Pemilu 2019 Pada Media Sosial Twitter Menggunakan Naive Bayes

Analisis Sentimen Persepsi Masyarakat Terhadap Pemilu 2019 Pada Media Sosial Twitter Menggunakan Naive Bayes
复制标题

分析 Sentimen Persepsi Masyarakat Terhadap Pemilu 2019 Pada Media Sosial Twitter Minggunakan Naive Bayes

DOI:
10.30865/mib.v4i3.2140
复制
发表时间:
2020
期刊:
JURNAL MEDIA INFORMATIKA BUDIDARMA
影响因子:
--
通讯作者:
Safitri Juanita
Safitri Juanita
中科院分区:
--
文献类型:
--
作者:
Safitri Juanita

文献摘要

参考文献

被引文献

相似文献

根据Bawaslu的评估,各种相关的负面内容支持潜在情侣涌入各种社交媒体页面。因此,有时内容会导致一个恶作剧问题,宗教和群体间种族(SARA)的问题。印尼人使用的社交媒体之一是推特,据Kompas.com报道,推特全球每日用户数量声称正在增加,这似乎是2019年第三季度推特财务报告在推特2019年第三季度财务报告中,推特平台上的日活跃用户记录增长了17%,达到1.45亿用户。因此,有必要进行情绪分析研究,捕捉社区在社交媒体推特上对2019年选举的看法模式,预计这项研究可以帮助感兴趣的各方在未来5年提高选民参与率。该研究方法使用了2018年4月16日至2019年4月16日的印度尼西亚推文数据,进一步的数据包括预处理、文本转换、词干印度尼西亚语、指定属性类、加载词典以及使用Weka对朴素贝叶斯进行分类。本研究的结论是对朴素贝叶斯进行分类,发现2019年选举推文数据集的负面感知模式为52%,远大于正面感知的18%,中性感知的值高于正面感知的31%。朴素贝叶斯对训练数据集的分类准确率为81%,对数据集的测试准确率为76%,对正面情绪的平均准确率为86.65%,对负面情绪的平均准确率为77.15%,对中性情绪的平均准确率为80.95%,对正面情绪的平均召回率为36.8%,对负面情绪的平均召回率为93.2%,对中性情绪的平均召回率为86.8%
According to the BAWASLU evaluation a variety of related negative content supports supporting prospective couples to burst into various social media pages. So sometimes the content leads to a hoax issue to the issue of religious and inter-group Racial (SARA). One of the social media used by the people of Indonesia is Twitter, according to Kompas.com number of Twitter daily users globally claimed to be increasing, this appears to be the 3rd Quarter Twitter Financial Report of 2019 on Twitter's 3rd quarter of 2019 Financial reports, daily active users on the Twitter platform are recorded to increase by 17 percent, to the number of 145 million users. So it is necessary that a sentiment analysis study can capture a pattern of community perception on social media Twitter against the 2019 elections and it is expected that this research can help interested parties to increase voter participation rate in the next 5 years. This research method uses the Indonesian tweet data taken from 16 April 2018-16 April 2019, further data in preprocessing, text transformation, stemming Bahasa Indonesia, specifying attribute class, load dictonary and a classification of Naive Bayes using Weka. The conclusion of this study was the classification of Naive Bayes finding that the 2019 election tweet dataset had a negative perception pattern of 52% much greater than the positive perception of 18% and the neutral perception had a value of 31% higher than positive perception. Naive Bayes ' degree of classification accuracy against the training dataset is 81% and the dataset testing 76%, the average precision value for positive sentiment is 86.65%, negative sentiment is 77.15%, and neutral sentiment is worth 80.95% while the average recall rate on positive sentiment is 36.8%, negative sentiment is 93.2% and the neutral sentiment is 86.8%
DOI: --
发表时间: 2014
期刊: Journal of management science
影响因子: --
作者:
อนิรุธ สืบสิงห์
通讯作者: อนิรุธ สืบสิงห์