Analisis Sentimen Dan Niat Pengguna Penyedia Layanan Internet Di Indonesia Menggunakan Metode Support Vector Machine

Nurqoulby, Fachrurrozy (2024) Analisis Sentimen Dan Niat Pengguna Penyedia Layanan Internet Di Indonesia Menggunakan Metode Support Vector Machine. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 17082010037-cover.pdf Download (1MB)
	Text (Bab 1) 17082010037-bab1.pdf Download (97kB)
	Text (Bab 2) 17082010037-bab2.pdf Restricted to Registered users only until 4 June 2026. Download (317kB)
	Text (Bab 3) 17082010037-bab3.pdf Restricted to Registered users only until 4 June 2026. Download (108kB)
	Text (Bab 4) 17082010037-bab4.pdf Restricted to Registered users only until 4 June 2026. Download (1MB)
	Text (Bab 5) 17082010037-bab5.pdf Download (11kB)
	Text (Daftar Pustaka) 17082010037-daftarpustaka.pdf Download (144kB)
	Text (Lampiran) 17082010037-lampiran.pdf Restricted to Registered users only until 4 June 2026. Download (542kB)

Abstract

In today's digital era, social media has become a major platform for individuals, especially on Twitter, to share opinions on various topics, including internet service providers. Sentiment analysis can identify a person's opinion expressed in text form. However, sentiment analysis is less able to reveal the motives and intentions of an opinion. Intention analysis can identify the motives, goals, and intentions behind a person's opinion. By combining sentiment and intention analysis, a deeper understanding can be obtained which can help in decision making. The stages required for sentiment and intention analysis are problem identification, literature study, needs analysis, model design, model evaluation, visualization, and report preparation. There are 7500 pieces of data with 3 classes for sentiment negative, neutral, and positive. As for the intention classes used, namely the complain, inquire, sell, praise, direct, compare, purchase, quit, criticize, and wish classes. However, there is an imbalance in the amount of data between classes. Therefore, SMOTE is needed, which is one method that can be used to balance the amount of data. The results of this study indicate that the SVM model with a Linear kernel obtains the highest accuracy of 83% without the SMOTE process compared to the RBF kernel of 82% and the polynomial of 75% for the sentiment model. In the intention model, SVM classification with Linear kernel also obtains the highest accuracy of 65% without the SMOTE process compared to the RBF kernel of 62% and the polynomial of 56%.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Kartika, Dhian Satria Yudha	201198 60 522249	dhian.satria@upnjatim.ac.id
Thesis advisor	Arifiyanti, Amalia Anjani	19920812 2018032 001	amalia_anjani.fik@upnjatim.ac.id

Subjects:

T Technology > T Technology (General) > T58.6-58.62 Management Information Systems

Divisions:

Faculty of Computer Science > Departemen of Information Systems

Depositing User:

Fachrurrozy Nurqoulby

Date Deposited:

05 Jun 2024 01:59

Last Modified:

20 Jun 2024 03:35

URI:

https://repository.upnjatim.ac.id/id/eprint/24299