ANALISIS SENTIMEN TERHADAP APLIKASI PEDULI LINDUNGI PADA JEJARING SOSIAL TWITTER MENGGUNAKAN ALGORITMA NAÏVE BAYES DAN SELEKSI FITUR PARTICLE SWARM OPTIMIZATION

• প্রধান লেখক: Muhamad Hanif Razka, (লেখক)
• বিন্যাস: গ্রন্থ
• প্রকাশিত: 2022-07-21.
• বিষয়গুলি: Thesis; NonPeerReviewed
• অনলাইন ব্যবহার করুন: Link Metadata
• Notes: http://repository.upnvj.ac.id/19792/1/ABSTRAK.pdf; http://repository.upnvj.ac.id/19792/2/AWAL.pdf; http://repository.upnvj.ac.id/19792/3/BAB%201.pdf; http://repository.upnvj.ac.id/19792/4/BAB%202.pdf; http://repository.upnvj.ac.id/19792/5/BAB%203.pdf; http://repository.upnvj.ac.id/19792/6/BAB%204.pdf; http://repository.upnvj.ac.id/19792/7/BAB%205.pdf; http://repository.upnvj.ac.id/19792/8/DAFTAR%20PUSTAKA.pdf; http://repository.upnvj.ac.id/19792/9/RIWAYAT%20HIDUP.pdf; http://repository.upnvj.ac.id/19792/10/LAMPIRAN.pdf; http://repository.upnvj.ac.id/19792/11/Hasil%20Plagiarisme.pdf; http://repository.upnvj.ac.id/19792/12/Artikel%20KI.pdf
• Keywords: QA Mathematics; QA75 Electronic computers. Computer science; QA76 Computer software; T Technology (General)

The concern protecting application is an official comprehensive application in collaboration with several other ministries. The application is intended to help prevent the spread of the covid-19 virus because it is constantly linked to its user by location. The study USES the public opinion data against the use of the care application of the tweets by people using key words such as protection, hashtags # protect and user username @plprotect. Data retrieval takes place March 13 to April 11, 2022. In this study the goal of categorizing a data of a tweet to a positive and negative sentiment and using a naive bayes algorithm to assess it and then apply the usage of personalizing personization of selection features by artificially matrix to further account for the accuracy of using the selection feature for the classification algorithm. And from the test results using the naive bayes classification algorithm, I was able to maintain a fixed rate of 76.78%, recall of 78%, and a precission of 79.62%. While the use of information regarding the swarm metrics optimization on algorithm naive gets the best results on pso's iteration process 250 times with an increased value of accuracy to 80.19% then recall value to 85.71% and also for precission to 80%.