ASPECT-BASED SENTIMENT ANALYSIS PADA ULASAN APLIKASI ACCESS BY KAI DENGAN MENGGUNAKAN METODE TF-IDF DAN ALGORITMA SUPPORT VECTOR MACHINE

Suryono, Muhammad Nur Rachman Nidhi (2025) ASPECT-BASED SENTIMENT ANALYSIS PADA ULASAN APLIKASI ACCESS BY KAI DENGAN MENGGUNAKAN METODE TF-IDF DAN ALGORITMA SUPPORT VECTOR MACHINE. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21082010137-cover.pdf Download (2MB)
	Text (Bab 1) 21082010137-bab1.pdf Download (88kB)
	Text (Bab 2) 21082010137-bab2.pdf Restricted to Repository staff only until 18 June 2027. Download (131kB) \| Request a copy
	Text (Bab 3) 21082010137-bab3.pdf Restricted to Repository staff only until 18 June 2027. Download (329kB) \| Request a copy
	Text (Bab 4) 21082010137-bab4.pdf Restricted to Repository staff only until 18 June 2027. Download (8MB) \| Request a copy
	Text (Bab 5) 21082010137-bab5.pdf Download (12kB)
	Text (Daftar Pustaka) 21082010137-daftarpustaka.pdf Download (148kB)

Abstract

Access by KAI is a digital transportation service application developed by PT Kereta Api Indonesia to simplify users' access to train travel services. To improve service quality and user experience, this study conducted sentiment analysis on user reviews from the Google Play Store using the Support Vector Machine (SVM) algorithm. The analysis focused on three key aspects: Financial Transactions, Technical Issues and Performance, and User Experience and Interface. The best result for Financial Transactions was achieved using the SMOTE model with a Linear kernel and a 70:30 data split, yielding an average f1-score of 0.66 (f1-score -1: 0.58, f1-score 0: 0.95, f1-score 1: 0.46). For Technical Issues and Performance, the highest performance came from the Non-SMOTE model with a Linear kernel and an 80:20 data split, resulting in an average f1-score of 0.76 (f1-score -1: 0.94, f1-score 0: 0.76, f1-score 1: 0.59). Meanwhile, the User Experience and Interface aspect achieved its best result with the SMOTE model using a Linear kernel and a 70:30 split, with an average f1-score of 0.76 (f1-score -1: 0.47, f1-score 0: 0.89, f1-score 1: 0.92). The best-performing model was then implemented into a Flask-based web application capable of predicting sentiment, exporting results in .csv format, and displaying data visualizations. This research aims to help PT KAI better understand user perceptions and enhance the Access by KAI service.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Arifiyanti, Amalia Anjani	NIDN0712089201	amalia_anjani.fik@upnjatim.ac.id
Thesis advisor	Kartika, Dhian Satria Yudha	NIDN0722058601	dhian.satria@upnjatim.ac.id

Subjects:

T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Information Systems

Depositing User:

Mr Suryono Muhammad Nur Rachman Nidhi

Date Deposited:

19 Jun 2025 04:34

Last Modified:

19 Jun 2025 04:34

URI:

https://repository.upnjatim.ac.id/id/eprint/38522