Analisis Sentimen Untuk Mengukur Kepuasan Pelanggan Terhadap Pelayanan Jasa Ekspedisi Tiki Dengan Metode Naive Bayes Classifier

Abdillah, Muhammad Ja’far (2019) Analisis Sentimen Untuk Mengukur Kepuasan Pelanggan Terhadap Pelayanan Jasa Ekspedisi Tiki Dengan Metode Naive Bayes Classifier. Undergraduate thesis, Politeknik Negeri Jember.

	Text (Abstrak) 7. ABSTRACT.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (263kB)
	Text (Bab 1 Pendahuluan) 14. BAB 1 PENDAHULUAN.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (173kB)
	Text (Daftar Pustaka) 19. DAFTAR PUSTAKA.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (207kB)
	Text (Laporan Lengkap) 21. E41151944_LAPORAN LENGKAP.pdf Restricted to Registered users only Download (2MB)

Abstract

According to data obtained from KOMINFO in 2014, internet users in Indonesia reached 82 million where 95% of these users were actively using social media. It caused influence in various fields of life, for example in this case the field of expedition services, namely the Tiki expedition company. Tiki also has social media in the form of Facebook Page to provide the latest info about the company. In the Facebook Page, customers provide comments or responses to the services provided by the company. These comments are very useful for knowing public opinion about the services provided by the company. Therefore, a study is needed to analyze public opinion on customer satisfaction with Tiki expedition services which can later be used as material for evaluation by Tiki. This research is analyzing the sentiments of comments or public opinions obtained on Facebook Page Tiki using Naïve Bayes Clasifier. Comment data is done by Text Processing to get data for the classification process. Then the data is classified into 3 categories, that is positive, neutral and negative using the Naïve Bayes Clasifier method. The results of the tests that have been carried out produce 83% system accuracy from the same dataset, 40% from different training and testing data with the same dataset, and 21% from different training data and data testing datasets.

Actions (login required)

View Item