ANALISIS SENTIMEN PADA MEDIA SOSIAL TWITTER MENGENAI KEBIJAKAN SELEKSI SEKOLAH JALUR ZONASI MENGGUNAKAN METODE KLASIFIKASI NAÏVE BAYES DAN SELEKSI FITUR PARTICLE SWARM OPTIMIZATION
Social media is a place to accommodate people's opinions or sentiments. An example of an application that is often used to discuss such sentiments is Twitter. Twitter users often express their opinions on several topics including education in particular government policies regarding zoning path...
Saved in:
Main Author: | |
---|---|
Format: | Book |
Published: |
2022-07-20.
|
Subjects: | |
Online Access: | Link Metadata |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Social media is a place to accommodate people's opinions or sentiments. An example of an application that is often used to discuss such sentiments is Twitter. Twitter users often express their opinions on several topics including education in particular government policies regarding zoning pathway school selection which can be viewed on Twitter. Based on the background of these conditions, research is needed on community sentiment regarding the zoning line school selection policy. This research was conducted using the Machine Learning method, namely the Naïve Bayes classification and feature selection using Particle Swarm Optimization to classify positive tweets or negative tweets that the community made, especially students and their parents who were fighting for seats in public schools. The results of this study prove that the Naïve Bayes classification with particle swarm Optimization feature selection gets an evaluation model value at greater accuracy than without using Particle Swarm Optimization where accuracy is 78%, precision is 23%, recall is 45%, and specificity by 82%. Meanwhile, the value of the Naïve Bayes classification evaluation mode without using the Particle Swarm Optimization feature selection is smaller, where the accuracy is 75%, the precision is 37%, the recall is 28%, and the specificity is 87%. There is an increase in performance in accuracy and recall, as well as a decrease in performance in precision and specificity. The evaluation value on precision and recall gets a low value, due to unbalanced data, where the ratio between positive and negative labels is 1: 4. |
---|---|
Item Description: | http://repository.upnvj.ac.id/19863/1/ABSTRAK.pdf http://repository.upnvj.ac.id/19863/13/AWAL.pdf http://repository.upnvj.ac.id/19863/3/BAB%201.pdf http://repository.upnvj.ac.id/19863/4/BAB%202.pdf http://repository.upnvj.ac.id/19863/5/BAB%203.pdf http://repository.upnvj.ac.id/19863/6/BAB%204.pdf http://repository.upnvj.ac.id/19863/7/BAB%205.pdf http://repository.upnvj.ac.id/19863/8/DAFTAR%20PUSTAKA.pdf http://repository.upnvj.ac.id/19863/9/DAFTAR%20RIWAYAT%20HIDUP.pdf http://repository.upnvj.ac.id/19863/10/LAMPIRAN.pdf http://repository.upnvj.ac.id/19863/11/HASIL%20PLAGIARISME.pdf http://repository.upnvj.ac.id/19863/12/ARTIKEL%20KI.pdf |