TY - GEN
T1 - Mining Indonesia Tourism's Reviews to Evaluate the Services through Multilabel Classification and LDA
AU - Laily, Irma Latifatul
AU - Budi, Indra
AU - Santoso, Aris Budi
AU - Putra, Prabu Kresna
N1 - Publisher Copyright:
© 2020 IEEE.
Copyright:
Copyright 2021 Elsevier B.V., All rights reserved.
PY - 2020/10/27
Y1 - 2020/10/27
N2 - The tourism sector is one of the mainstay factors, one of the most significant economic contributors in Lamongan. There are two leading tourism destinations in Lamongan, namely WBL and Mazoola.Evaluation of tourist experience in a tourist destination can use the reviews provided at the end of the trip. Tourists review various aspects of tourism, such as price, services, and location. Classify more than one aspect from reviews is a challenging task. Five labels used, namely: Price, Location, Safety, Services and Facilities, and Environment and Ambiance. This study was conducted to determine the aspects that should be evaluated from the reviews that visitors provide. This research uses five multi-label classifier algorithms commonly used for multi-label classification: NBSVM, Binary Relevance-Naive Bayes, Binary Relevance-Logistic Regression, Classifier Chains-Naive Bayes, and Multilabel kNN. NBSVM was a robust performer. For WBL data in scenario 1, the highest accuracy isBR-LR, which is 92%. Whereas in scenario 2, NBSVM has the highest value of 91,32%. However, in other assessments, NBSVM is still superior. Likewise to Mazola's data, NBSVM has the highest accuracy in both scenarios: 87,38% and 87,28%. This study also extracts three trend topics for each data set,WBL and Mazoola. Trend topics aim to find out what topics are discussed more frequently in each tourist destination review-topic extraction using LDA with the Gensim library.
AB - The tourism sector is one of the mainstay factors, one of the most significant economic contributors in Lamongan. There are two leading tourism destinations in Lamongan, namely WBL and Mazoola.Evaluation of tourist experience in a tourist destination can use the reviews provided at the end of the trip. Tourists review various aspects of tourism, such as price, services, and location. Classify more than one aspect from reviews is a challenging task. Five labels used, namely: Price, Location, Safety, Services and Facilities, and Environment and Ambiance. This study was conducted to determine the aspects that should be evaluated from the reviews that visitors provide. This research uses five multi-label classifier algorithms commonly used for multi-label classification: NBSVM, Binary Relevance-Naive Bayes, Binary Relevance-Logistic Regression, Classifier Chains-Naive Bayes, and Multilabel kNN. NBSVM was a robust performer. For WBL data in scenario 1, the highest accuracy isBR-LR, which is 92%. Whereas in scenario 2, NBSVM has the highest value of 91,32%. However, in other assessments, NBSVM is still superior. Likewise to Mazola's data, NBSVM has the highest accuracy in both scenarios: 87,38% and 87,28%. This study also extracts three trend topics for each data set,WBL and Mazoola. Trend topics aim to find out what topics are discussed more frequently in each tourist destination review-topic extraction using LDA with the Gensim library.
KW - LDA
KW - Multi/abel Classification
KW - Tourism Review
UR - http://www.scopus.com/inward/record.url?scp=85100278216&partnerID=8YFLogxK
U2 - 10.1109/ICELTICs50595.2020.9315392
DO - 10.1109/ICELTICs50595.2020.9315392
M3 - Conference contribution
AN - SCOPUS:85100278216
T3 - Proceedings of the International Conference on Electrical Engineering and Informatics
BT - 2020 International Conference on Electrical Engineering and Informatics, ICELTICs 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 International Conference on Electrical Engineering and Informatics, ICELTICs 2020
Y2 - 27 October 2020 through 28 October 2020
ER -