Siroj, Sobakhul Munir (2021) Analisis Sentimen Opini Publik pada Twitter Terhadap Efek Pembelajaran Daring di Universitas Brawijaya Menggunakan Metode K-Nearest Neighbor. Sarjana thesis, Universitas Brawijaya.
Abstract
Akibat pandemi virus Covid-19, Indonesia menerapkan sistem pembelajaran daring, termasuk di Universitas Brawijaya. Pelaksanaan sistem pembelajaran daring di Universitas Brawijaya menuai banyak pendapat positif dan pendapat negatif, dan bahkan sempat menjadi trending topic di media sosial twitter pada akhir bulan April tahun 2020 lalu. Penelitian ini bertujuan untuk mengetahui tingkat kecenderungan pendapat positif dan pendapat negatif melalui proses klasifikasi sentimen, serta mengetahui beberapa hal yang sering dijadikan keluhan dalam pelaksanaan sistem pembelajaran daring di Universitas Brawijaya. Proses klasifikasi sentimen dilakukan dengan menggunakan metode K-Nearest Neighbor pada aplikasi RapidMiner. Metode K-Nearest Neighbor dipilih karena memiliki prinsip sederhana, mudah digunakan, serta cukup baik digunakan untuk proses klasifikasi sentimen. Proses klasifikasi sentimen pada penelitian ini dilakukan melalui empat tahapan utama, yaitu tahap pengumpulan data, tahap pra-pemrosesan data dan pembobotan term, tahap klasifikasi, dan tahap pengujian. Selain itu, dilakukan juga tahap perhitungan frekuensi kemunculan kata secara terpisah. Proses klasifikasi sentimen dilakukan terhadap 200 data latih dan 50 data uji. Dari proses klasifikasi didapatkan hasil bahwa 50,8% memberikan pendapat negatif, sedangkan sisanya sebesar 49,2% memberikan pendapat positif terhadap pelaksanaan sistem pembelajaran daring di Universitas Brawijaya. Kemudian dari proses perhitungan frekuensi kemunculan kata, didapatkan juga lima kata yang sering muncul dan banyak dikeluhkan, yaitu kata “offline”, “dosen”, “tugas”, “kuota”, dan “ukt”. Kelima kata tersebut secara berturut-turut memiliki presentase sebesar 6%, 6,4%, 5,2%, 4,8%, dan 3,2%. Pada proses pengujian, dilakukan dengan memberikan variasi nilai k dan pengaruhnya terhadap nilai accuracy, precision, dan recall. Selain itu juga dilakukan proses analisis pengujian, cross validation, dan tahap feedback. Dari hasil pengujian, didapatkan hasil bahwa nilai accuracy, precision, dan recall terbaik secara berturut-turut adalah 80% untuk nilai k=7, 81,48% untuk nilai k=7, dan 88,89% untuk nilai k=23.
English Abstract
As a result of the Covid-19 virus pandemic, Indonesia has implemented an online learning system, including at Brawijaya University. The implementation of the online learning system at Brawijaya University has reaped many positive and negative opinions, and even became a trending topic on Twitter social media at the end of April 2020. This study aims to determine the level of tendency of positive opinions and negative opinions through the sentiment classification process, as well as knowing some things that are often used as complaints in the implementation of the online learning system at Brawijaya University. The sentiment classification process is carried out using the K-Nearest Neighbor method in the Rapid Miner application. The K-Nearest Neighbor method was chosen because it has simple principles, is easy to use, and is good enough to use for the sentiment classification process. The sentiment classification process in this study was carried out through four main stages, namely the data collection stage, the data pre-processing stage and term weighting, the classification stage, and the testing stage. In addition, a separate calculation of the frequency of occurrence of words is also carried out. The sentiment classification process was carried out on 200 training data and 50 test data. From the classification process, it was found that 50.8% gave negative opinions, while the remaining 49.2% gave positive opinions on the implementation of the online learning system at Brawijaya University. Then from the process of calculating the frequency of occurrence of words, there were also five words that often appeared and complained about a lot, namely the word "offline", "dosen", "tugas", "kuota", and "ukt". The five words respectively have a percentage of 6%, 6,4%, 5,2%, 4,8%, and 3,2%. In the testing process, it is carried out by providing variations in the k value and its effect on the accuracy, precision, and recall values. In addition, the testing analysis process, cross validation, and feedback stages are also carried out. From the test results, it was found that the best accuracy, precision, and recall values were 80% for the value of k = 7, 81,48% for the value of k = 7, and 88,89% for the value of k = 23, respectively.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | 0521150041 |
| Uncontrolled Keywords: | belajar daring, k-nearest neighbor, klasifikasi sentimen, rapidminer, twitter, universitas brawijaya, online learning, k-nearest neighbor, sentiment classification, rapidminer, twitter, brawijaya university |
| Subjects: | 000 Computer science, information and general works > 004 Computer science |
| Divisions: | Fakultas Ilmu Komputer > Sistem Informasi |
| Depositing User: | Budi Wahyono Wahyono |
| Date Deposited: | 01 Nov 2021 01:48 |
| Last Modified: | 01 Oct 2024 07:39 |
| URI: | http://repository.ub.ac.id/id/eprint/186233 |
![[thumbnail of Sobakhul Munir Siroj.pdf]](http://repository.ub.ac.id/style/images/fileicons/text.png) |
Text
Sobakhul Munir Siroj.pdf Download (2MB) |
Actions (login required)
 |
View Item |