Analisis Sentimen Komentar Masyarakat Terhadap Jetski Mahakam Pada Media Sosial TikTok Menggunakan Pendekatan Lexicon

Atifah, Siti Nur (2025) Analisis Sentimen Komentar Masyarakat Terhadap Jetski Mahakam Pada Media Sosial TikTok Menggunakan Pendekatan Lexicon. S1 Sistem Informasi thesis, STMIK Widya Cipta Dharma.

[img] Text
2141010-S1- Sistem Informasi.pdf
Restricted to Repository staff only
Download (2MB) | Request a copy
[img] Text
2141010-S1- Jurnal.pdf
Download (580kB)

Abstract

Penelitian ini bertujuan untuk melakukan analisis sentimen terhadap komentar masyarakat pada media sosial TikTok terkait wisata Jetski Mahakam. Analisis ini dilakukan guna mengetahui persepsi publik yang terekam melalui komentar-komentar pengguna, serta sebagai bahan evaluasi untuk meningkatkan kualitas layanan dan strategi promosi wisata. Penelitian ini dilakukan pada wisata Jetski Mahakam di Samarinda. Teknik pengumpulan data yang digunakan meliputi wawancara dengan pihak pengelola, studi pustaka yang mengkaji literatur relevan, serta studi lapangan dengan melakukan pengumpulan data komentar secara langsung dari akun TikTok Jetski Mahakam menggunakan metode Web scraping dengan bantuan platform Apify. Dalam penelitian ini digunakan pendekatan Lexicon dengan memanfaatkan kamus InSet (Indonesia Sentiment Lexicon). Proses penelitian mengikuti tahapan CRISP-DM, yaitu business understanding, data understanding, data preparation, modelling, Evaluation, dan deployment. Komentar yang berhasil dikumpulkan sebanyak 505, dan setelah proses pembersihan data tersisa 422 komentar yang dianalisis. Setiap komentar diklasifikasikan ke dalam kategori sentimen positif, negatif, atau netral, baik secara otomatis menggunakan kamus Lexicon maupun secara manual. Hasil penelitian menunjukkan Hasil Labelling manual menghasilkan 85 sentimen positif, 223 sentimen negatif, dan 114 sentimen netral, hasil Labelling menggunakan kamus Lexicon InSet menghasilkan 60 sentimen positif, 290 sentimen negatif, dan 72 sentimen netral, dan Labelling GPT menghasilkan 29 sentimen positif, 38 sentimen negatif, dan 355 sentimen netral. Pada tahap model Evaluation dihasilkan perbandingan akurasi yaitu, perbandingan Labelling manual dan Labelling InSet menghasilkan akurasi 71.56%, perbandingan Labelling Manual dan Labelling GPT menghasilkan akurasi 41.70%, perbandingan Labelling InSet dan Labelling GPT menghasilkan akurasi 28.43%%. Dari hasil tersebut dapat disimpulkan bahwa pendekatan Lexicon menggunakan kamus InSet cukup efektif untuk analisis sentimen. Visualisasi di sistem dalam bentuk pie chart dan word cloud yang memudahkan pemahaman distribusi sentimen serta kata-kata dominan dalam komentar. ============================================================ This study aims to conduct sentiment analysis on public comments on TikTok related to the Jetski Mahakam tourism activity. The analysis was carried out to understand public perception as reflected in user comments and to serve as an Evaluation tool for improving service quality and tourism promotion strategies.The research was conducted on the Jetski Mahakam tourism activity in Samarinda. Data collection techniques included interviews with the management, literature Reviews of relevant studies, and field studies by directly collecting comment data from the Jetski Mahakam TikTok account using Web scraping methods with the help of the Apify platform.This study employed a Lexicon-Based approach using the InSet (Indonesia Sentiment Lexicon) dictionary. The research process followed the CRISP-DM stages, namely business understanding, data understanding, data preparation, modelling, Evaluation, and deployment. A total of 505 comments were collected, and after data cleaning, 422 comments were analyzed. Each comment was classified into positive, negative, or neutral sentiment categories, either automatically using the Lexicon or manually.The research findings showed that the Manual Labelling produced 85 positive sentiments, 223 negative sentiments, and 114 neutral sentiments, Lexicon-Based Labelling using the InSet dictionary resulted in 60 positive sentiments, 290 negative sentiments, and 72 neutral sentiments. Meanwhile, GPT-Based Labelling generated 29 positive sentiments, 38 negative sentiments, and 355 neutral sentiments. In the model Evaluation phase, the comparison between manual and InSet Labelling yielded an accuracy of 71.56%, Manual and GPT Labelling showed 41.70%% accuracy, and InSet and GPT Labelling resulted in 28.43% accuracy. These results indicate that the Lexicon-Based approach using the InSet dictionary is fairly effective for sentiment analysis. The system visualizes the results in the form of pie charts and word clouds, which help in understanding the distribution of sentiments and the dominant words in the comments.

Item Type: Thesis (S1 Sistem Informasi)
Additional Information: Pembimbing 1 : Wahyuni, S.Kom., M.Kom Pembimbong 2 : Muhammad Ibnu Sa'ad, S.Kom., M.Kom
Uncontrolled Keywords: Analisis Sentimen, Lexicon, TikTok, InSet, CRISP-DM, Jetski Mahakam
Subjects: Q Science > QA Mathematics > QA76 Computer software
Divisions: Sistem Informasi
Depositing User: Ms Siti Nur Atifah
Date Deposited: 07 Aug 2025 06:36
Last Modified: 07 Aug 2025 06:36
URI: http://repository.wicida.ac.id/id/eprint/6212

Actions (login required)

View Item View Item
wicida repository is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.