Page 50 - Conference Book The 3rd IConVET "Future of TVET Graduates: Developing Talent for Industry 4.0 and The New Normal"

The 3 International Conference
rd
on Vocational Education and Technology (IConVET)
Universitas Pendidikan Ganesha - Bali | 7 November, 2020
th
SENTIMENT ANALYSIS OF ONLINE TRAVEL AGENT
USING NAÏVE BAYES AND K-NEAREST NEIGHBOR

Selviana Yunita and Eka W Sholeha 2
1

1 Darwan Ali University, Sampit, Indonesia
2 Politeknik Negeri Tanah Laut, Pelaihari, Indonesia

1 selviana.yunita.ax@gmail.com, ekaaws8@gmail.com
2

Abstract. Social media has impact for decision maker to get more insights
broadly. Including for online travel agent company, where costumer’s
interest to use online travel agent for their chosen agent will grows along
with the high number of customer’s satisfaction. As a one of the most
important point in distribution, company provides a platform that reliable
and effective to purchase a trip and share information of their experience
through Online travel agent. It is important to know how consumer
considerate which one the online travel agent they choose. One of their
method is looking at the reviews. Facebook is one of social media that
provide numerous reviews through comments sections. The research
purposes are twofold, algorithm comparison and reveal the effect of
uppercase as well as punctuation mark. The accuracy comparison between
Naïve Bayes and K-Nearest Neighbor is provided against the datasets. This
research collects the data from user comments on facebook about the biggest
three online travel agents in Indonesia. We classify the comments into three
categories which are positive, negative, and neutral. The result of this
research is found that K-Nearest Neighbor have slightly higher accuracy
than the Naïve Bayes. Moreover, lowercase text without punctuation
achieves better accuracy for both of algorithm.

| IConVET 2020 | 50

https://conference.undiksha.ac.id/iconvet/ | 50 |