PERDANA, Rheza Akbar and Widodo, Catur Edi and Santoso, Rukun (2024) KOMPARASI ALGORITMA NAIVE BAYES, DECISION TREE, DAN K-NEAREST NEIGHBOR (K-NN) DALAM ANALISIS SENTIMEN CYBERBULLYING PADA KOMENTAR INSTAGRAM. Masters thesis, UNIVERSITAS DIPONEGORO.
|
Text
COVER-1 (1).pdf Download (14kB) |
|
|
Text
COVER_dll.pdf Restricted to Repository staff only Download (3MB) |
|
|
Text
BAB I.pdf Download (107kB) |
|
|
Text
BAB II.pdf Download (4MB) |
|
|
Text
BAB III.pdf Restricted to Repository staff only Download (4MB) |
|
|
Text
BAB IV.pdf Restricted to Repository staff only Download (961kB) |
|
|
Text
BAB V.pdf Restricted to Repository staff only Download (168kB) |
|
|
Text
DAFTAR PUSTAKA SEMINAR.pdf Download (242kB) |
|
|
Text
LAMPIRAN.pdf Restricted to Repository staff only Download (802kB) |
Abstract
Instagram merupakan media sosial yang paling populer pada zaman sekarang. Pengguna yang dimulai dari anak-anak, remaja hingga orang dewasa turut mendongkrak popularitas Instagram. Namun, media sosial ini tidak lepas dari bahaya cyberbullying yang sering dilakukan oleh pengguna khususnya pada kolom komentar. Dengan data statistik yang telah didapatkan, bahwa 42% remaja berusia 12-20 tahun telah menjadi korban cyberbullying. Bahaya cyberbullying tentunya meresahkan banyak orang dikarenakan dampak yang ditimbulkan, maka dari itu dapat dilakukan suatu analisis sentimen pada kolom komentar Instagram yang berupaya untuk mengetahui sentimen dari setiap komentar. Analisis sentimen merupakan suatu cabang ilmu dari text mining yang digunakan untuk mengekstrak, memahami, dan mengolah data teks. Untuk mengetahui setiap sentimen pada komentar digunakan metode klasifikasi Naïve Bayes, Decision Tree, dan K-Nearest Neighbor (K-NN) selanjutnya digunakan Teknik Undersampling untuk menyeimbangkan data sentimen yang memiliki jumlah kelas sentimen lebih sedikit. Analisis sentimen dengan melakukan crawling pada data instagram sebanyak 800 data komentar yang berbahasa Indonesia menggunakan python dengan tools Google Colab. Dan hasil pengklasifikasikan tersebut menghasilkan nilai akurasi untuk masing-masing metode berturut-turut adalah berdasarkan menggunakan Teknik Undersampling 89%, 79%, dan 75% sedangkan berdasarkan menggunakan Teknik Non Undersampling 88%, 82%, dan 75% untuk data komentar tersebut. Nilai akurasi tersebut menunjukkan bahwa metode Naïve Bayes paling baik untuk mengklasifikasikan data komentar Instagram berdasarkan menggunakan Teknik Undersampling dan Non Undersampling.
Kata kunci : Cyberbullying, Analisis Sentimen, Naive Bayes, Decision Tree, K-Nearest Neighbor (K-NN)
Instagram is the most popular social media today. Users ranging from children, teenagers to adults have also boosted Instagram's popularity. However, this social media is not free from the dangers of cyberbullying that are often carried out by users, especially in the comments column. With the statistical data that has been obtained, 42% of teenagers aged 12-20 years have become victims of cyberbullying. The dangers of cyberbullying are certainly disturbing to many people because of the impacts caused, therefore a sentiment analysis can be carried out on the Instagram comment column which attempts to find out the sentiment of each comment. Sentiment analysis is a branch of text mining that is used to extract, understand, and process text data. To find out each sentiment in the comments, the Naïve Bayes, Decision Tree, and K-Nearest Neighbor (K-NN) classification methods are used, then the Undersampling Technique is used to balance sentiment data that has a smaller number of sentiment classes. Sentiment analysis by crawling Instagram data of 800 Indonesian-language comment data using python with Google Colab tools. And the results of the classification produce accuracy values for each method respectively based on using the Undersampling Technique 89%, 79%, and 75% while based on using the Non-Undersampling Technique 88%, 82%, and 75% for the comment data. The accuracy value shows that the Naïve Bayes method is best for classifying Instagram comment data based on using the Undersampling and Non-Undersampling Techniques.
Keywords : Cyberbullying, Sentiment Analysis, Naïve Bayes, Decision Tree, K-Nearest Neighbor (K-NN)
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Cyberbullying, Analisis Sentimen, Naive Bayes, Decision Tree, K-Nearest Neighbor (K-NN) |
| Subjects: | Sciences and Mathemathic |
| Divisions: | Postgraduate Program > Master Program in Information System |
| Depositing User: | ekana listianawati |
| Date Deposited: | 15 May 2025 08:37 |
| Last Modified: | 15 May 2025 08:37 |
| URI: | https://eprints2.undip.ac.id/id/eprint/32244 |
Actions (login required)
![]() |
View Item |
