OPTIMASI ALGORITMA SUPPORT VECTOR MACHINE MENGGUNAKAN SELEKSI FITUR PARTICLE SWARM OPTIMIZATION PADA ANALISIS SENTIMEN TERHADAP KEBIJAKAN PPKM

• Main Author: Hasan Mubarok, (Author)
• Format: Book
• Published: 2022-06-07.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upnvj.ac.id/19754/1/ABSTRAK.pdf; http://repository.upnvj.ac.id/19754/2/AWAL.pdf; http://repository.upnvj.ac.id/19754/3/BAB%201.pdf; http://repository.upnvj.ac.id/19754/4/BAB%202.pdf; http://repository.upnvj.ac.id/19754/5/BAB%203.pdf; http://repository.upnvj.ac.id/19754/6/BAB%204.pdf; http://repository.upnvj.ac.id/19754/7/BAB%205.pdf; http://repository.upnvj.ac.id/19754/8/DAFTAR%20PUSTAKA.pdf; http://repository.upnvj.ac.id/19754/9/RIWAYAT%20HIDUP.pdf; http://repository.upnvj.ac.id/19754/10/LAMPIRAN.pdf; http://repository.upnvj.ac.id/19754/11/HASIL%20PLAGIARISME.pdf; http://repository.upnvj.ac.id/19754/12/ARTIKEL%20KI.pdf
• Keywords: QA75 Electronic computers. Computer science; QA76 Computer software; T Technology (General)

Twitter is a micro-blogging social media that allows users to express opinions on various topics and discuss current issues. One of the topics that are often discussed by the public is the implementation of PPKM in Indonesia which raises pros and cons so that opinions from the public are very diverse, especially Twitter users. With the number of opinions, it is necessary to have a sentiment analysis. The aim is to find out public opinion on the implementation of PPKM through the hashtag #PPKM. Therefore, this research carried out the classification process on the application of PPKM using two classes, namely the positive sentiment class and the negative sentiment class. The method used in classifying is the Support Vector Machine algorithm and the Particle Swarm Optimization algorithm as feature selection. The two algorithms will be divided into two processes, namely making models using PSO and not using PSO. Retrieval by crawling technique starts from July 1 - August 30, 2021, with the API that has been provided by Twitter. The results of the classification evaluation using the confusion matrix obtained an accuracy value of 79.77%, recall 69.04%, and 85.29% on data without PSO (Feature Selection). While the data using PSO (Feature Selection) obtained an accuracy value of 87.08%, recall 76.83%, and Precision 94.03%.