Putri, Siti Oktavia Eka (2025) ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN TEMPAT WISATA POPULER DI JAWA TIMUR MENGGUNAKAN CNN. Undergraduate thesis, UPN Veteran Jawa Timur.
This is the latest version of this item.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
21082010081-cover.pdf Download (7MB) |
|
Text (Bab 1)
21082010081-bab1.pdf Download (420kB) |
|
Text (Bab 2)
21082010081-bab2.pdf Restricted to Repository staff only until 30 June 2027. Download (656kB) |
|
Text (Bab 3)
21082010081-bab3.pdf Restricted to Repository staff only until 30 June 2027. Download (552kB) |
|
Text (Bab 4)
21082010081-bab4.pdf Restricted to Repository staff only until 30 June 2027. Download (6MB) |
|
Text (Bab 5)
21082010081-bab5.pdf Download (248kB) |
|
Text (Daftar Pustaka)
21082010081-daftarpustaka.pdf Download (231kB) |
|
Text (Lampiran)
21082010081-lampiran.pdf Restricted to Repository staff only Download (220kB) |
Abstract
The tourism industry is one of the key sectors in Indonesia's economy, showing a positive recovery trend in the post-COVID-19 period. East Java, as the province with the highest number of tourist visits in Indonesia, holds great potential for further analysis to support efforts toward sustainable tourism recovery and development. This study aims to implement deep learning using a Convolutional Neural Network (CNN)-based algorithm to perform aspect-based sentiment prediction on reviews of popular tourist destinations in East Java and to identify the scenario that yields the best performance. The aspects used in this study are attraction, amenities, access, and price, with review data collected from digital platforms such as TripAdvisor and Google Maps. The method applied is Aspect Based Sentiment Analysis (ABSA) using a deep learning approach with CNN, integrated with different word embeddings, namely Word2Vec, BERT, and IndoBERT. A total of five models were built: one multilabel aspect classification model and four sentiment classification models for each aspect. The results show that the model using IndoBERT embedding provided the best performance with F1-score and AUC as follows: 1) The aspect classification model achieved 0.81 and 0.91, 2) The sentiment classification model for Attraction achieved 0.71 and 0.94, 3) For Amenities, 0.85 and 0.93, 4) For Access, 0.76 and 0.83, and 5) For Price, 0.81 and 0.89. Once the best-performing model was selected, it was implemented into a web-based application to perform aspect-based sentiment predictions on tourism reviews, capable of accepting both single review input and batch reviews via CSV files. The prediction results display sentiment for each detected aspect, and for batch reviews, a pie chart is also provided to facilitate the interpretation of the predicted data.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Information Systems | ||||||||||||
| Depositing User: | Siti Oktavia Eka Putri | ||||||||||||
| Date Deposited: | 30 Jun 2025 08:00 | ||||||||||||
| Last Modified: | 30 Jun 2025 08:00 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/39140 |
Available Versions of this Item
-
ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN TEMPAT WISATA POPULER DI JAWA TIMUR MENGGUNAKAN CNN. (deposited UNSPECIFIED)
- ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN TEMPAT WISATA POPULER DI JAWA TIMUR MENGGUNAKAN CNN. (deposited 30 Jun 2025 08:00) [Currently Displayed]
Actions (login required)
 |
View Item |