Putra, Ramzhy Deviannico (2025) Pengembangan Platform Booking Jasa Kursus Berselancar Berbasis Website untuk Optimalisasi Layanan dengan Ruby On Rails dan Analisa Sentimen. Undergraduate thesis, Politeknik Negeri Bali.

[thumbnail of Full Text] Text (Full Text)
RAMA_58302_2115354078_full.pdf - Accepted Version
Restricted to Repository staff only
Download (7MB) | Request a copy
[thumbnail of iThenticate] Text (iThenticate)
RAMA_58302_2115354078_iThenticate.pdf - Accepted Version
Restricted to Repository staff only
Download (1MB) | Request a copy
[thumbnail of Cover, Bab 1, Bab 5 dan Referensi] Text (Cover, Bab 1, Bab 5 dan Referensi)
RAMA_58302_2115354078_004097606_0817128601_Part.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (279kB)

Abstract

In the digital era, Natural Language Processing (NLP) plays a crucial role in understanding user opinions about a service. This research aims to develop an automated sentiment analysis system integrated into a web-based booking platform for surfing course services. The system is designed to automatically process customer feedback written in either Japanese or English. The NLP pipeline begins with language detection using langdetect, followed by automatic translation via GoogleTranslator for texts in Japanese. Next, spelling and grammar correction is performed using TextBlob, before analyzing sentiment polarity through a lexicon-based approach. Negative reviews are grouped into major topics using TF-IDF and KMeans clustering methods. Each negative cluster is then analyzed using GPT (OpenAI) to generate automatic improvement suggestions. This structured NLP pipeline enables the system to effectively handle multilingual feedback and provide meaningful and actionable insights to support decision-making in service quality enhancement. The results demonstrate that integrating NLP and AI into a digital service system can optimize customer experience management by delivering real-time sentiment analysis and data-driven recommendations. This approach holds significant potential for broader implementation in various digital service sectors.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Natural Language Processing, Sentiment Analysis, Lexicon-Based, TF-IDF, KMeans
Subjects: Ilmu Teknik > Teknik Elektro Dan Informatika > Sistem Informasi
Ilmu Teknik > Teknik Elektro Dan Informatika > Teknologi Informasi
Divisions: Jurusan Teknologi Informasi > Prodi D4 Teknologi Rekayasa Perangkat Lunak > Skripsi
Depositing User: Ramzhy Deviannico Putra
Date Deposited: 19 Sep 2025 03:10
Last Modified: 19 Sep 2025 03:10
URI: https://repository.pnb.ac.id/id/eprint/17589

Actions (login required)

View Item View Item