<abstract xmlns="http://www.w3.org/1999/xhtml"><sec id="j_jdis-2019-0005_s_006_w2aab3b7b5b1b6b1aab1c17b1Aa"><h3>Purpose</h3><p>Online reviews on tourism attractions provide important references for potential tourists to choose tourism spots. The main goal of this study is conducting sentiment analysis to facilitate users comprehending the large scale of the reviews, based on the comments about Chinese attractions from Japanese tourism website 4Travel.</p></sec><sec id="j_jdis-2019-0005_s_007_w2aab3b7b5b1b6b1aab1c17b2Aa"><h3>Design/methodology/approach</h3><p>Different statistics- and rule-based methods are used to analyze the sentiment of the reviews. Three groups of novel statistics-based methods combining feature selection functions and the traditional term frequency-inverse document frequency (TF-IDF) method are proposed. We also make seven groups of different rules-based methods. The macro-average and micro-average values for the best classification results of the methods are calculated respectively and the performance of the methods are shown.</p></sec><sec id="j_jdis-2019-0005_s_008_w2aab3b7b5b1b6b1aab1c17b3Aa"><h3>Findings</h3><p>We compare the statistics-based and rule-based methods separately and compare the overall performance of the two method. According to the results, it is concluded that the combination of feature selection functions and weightings can strongly improve the overall performance. The emotional vocabulary in the field of tourism (EVT), kaomojis, negative and transitional words can notably improve the performance in all of three categories. The rule-based methods outperform the statistics-based ones with a narrow advantage.</p></sec><sec id="j_jdis-2019-0005_s_009_w2aab3b7b5b1b6b1aab1c17b4Aa"><h3>Research limitation</h3><p>Two limitations can be addressed: 1) the empirical studies to verify the validity of the proposed methods are only conducted on Japanese languages; and 2) the deep learning technology is not been incorporated in the methods.</p></sec><sec id="j_jdis-2019-0005_s_010_w2aab3b7b5b1b6b1aab1c17b5Aa"><h3>Practical implications</h3><p>The results help to elucidate the intrinsic characteristics of the Japanese language and the influence on sentiment analysis. These findings also provide practical usage guidelines within the field of sentiment analysis of Japanese online tourism reviews.</p></sec><sec id="j_jdis-2019-0005_s_011_w2aab3b7b5b1b6b1aab1c17b6Aa"><h3>Originality/value</h3><p>Our research is of practicability. Currently, there are no studies that focus on the sentiment analysis of Japanese reviews about Chinese attractions.</p></sec></abstract>

PurposeOnline reviews on tourism attractions provide important references for potential tourists to choose tourism spots. The main goal of this study is conducting sentiment analysis to facilitate users comprehending the large scale of the reviews, based on the comments about Chinese attractions from Japanese tourism website 4Travel.Design/methodology/approachDifferent statistics- and rule-based methods are used to analyze the sentiment of the reviews. Three groups of novel statistics-based methods combining feature selection functions and the traditional term frequency-inverse document frequency (TF-IDF) method are proposed. We also make seven groups of different rules-based methods. The macro-average and micro-average values for the best classification results of the methods are calculated respectively and the performance of the methods are shown.FindingsWe compare the statistics-based and rule-based methods separately and compare the overall performance of the two method. According to the results, it is concluded that the combination of feature selection functions and weightings can strongly improve the overall performance. The emotional vocabulary in the field of tourism (EVT), kaomojis, negative and transitional words can notably improve the performance in all of three categories. The rule-based methods outperform the statistics-based ones with a narrow advantage.Research limitationTwo limitations can be addressed: 1) the empirical studies to verify the validity of the proposed methods are only conducted on Japanese languages; and 2) the deep learning technology is not been incorporated in the methods.Practical implicationsThe results help to elucidate the intrinsic characteristics of the Japanese language and the influence on sentiment analysis. These findings also provide practical usage guidelines within the field of sentiment analysis of Japanese online tourism reviews.Originality/valueOur research is of practicability. Currently, there are no studies that focus on the sentiment analysis of Japanese reviews about Chinese attractions.

Sentiment Analysis of Japanese Tourism Online Reviews

Journal of Data and Information Science

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

{"article-title":"Sentiment Analysis of Japanese Tourism Online Reviews"}

PurposeOnline reviews on tourism attractions provide important references for potential tourists to choose tourism spots. The main goal of this study is...

	macro-average	micro-average
Group 1	0.4606	0.5350
Group 2	0.6419	0.6891
Group 3	0.6436	0.6924
Group 4	0.6436	0.6924
Group 5	0.6422	0.6914
Group 6	0.6527	0.6940
Group 7	0.6847	0.7185

	Positive			Neutral			Negative
	Precision	Recall	F1	Precision	Recall	F1	Precision	Recall	F1
Group 1	0.8366	0.3282	0.4714	0.4596	0.8967	0.6077	0.4780	0.2216	0.3028
Group2	0.7691	0.7547	0.7618	0.6429	0.6494	0.6462	0.5056	0.5306	0.5178
Group3	0.7713	0.7618	0.7665	0.6478	0.6494	0.6486	0.5042	0.5277	0.5157
Group4	0.7713	0.7618	0.7665	0.6478	0.6494	0.6486	0.5042	0.5277	0.5157
Group 5	0.7713	0.7618	0.7665	0.6475	0.6469	0.6472	0.4986	0.5277	0.5127
Group6	0.7790	0.7508	0.7647	0.6426	0.6562	0.6493	0.5214	0.5685	0.5439
Group7	0.8057	0.7625	0.7835	0.6539	0.6926	0.6727	0.5871	0.6093	0.5980

	Macro-average	Micro-average
WE+SVM	0.2241	0.5065
WE+MLP	02924	0.5277
WE+LR	0.3573	0.5619
WE+NB	0.3595	0.4121
WE+DT	0.3996	0.4723
WE+RF	0.4094	0.5326
WE+LSTM	0.4746	0.5668
WE+BiLSTM	0.4831	0.5602
CHI*TF-IDF (Laplace)	0.6379	0.7020
Rule-based (Group7)	0.6847	0.7185

		DisI	DisII	DisIII
positive	precision	0.8057	0.8057	0.8057
	recall	0.7625	0.7625	0.7625
	F1	0.7835	0.7835	0.7835
neutral	precision	0.6539	0.6539	0.6539
	recall	0.6926	0.6943	0.6943
	F1	0.6727	0.6735	0.6735
negative	precision	0.5871	0.5921	0.5921
	recall	0.6093	0.6093	0.6093
	F1	0.5980	0.6006	0.6006
macro-average		0.6847	0.6857	0.6859
micro-average		0.7185	0.7191	0.7191

Sentiment Analysis of Japanese Tourism Online Reviews

Article Category: Research Paper

Data publikacji: 21 lut 2019

Zakres stron: 89 - 113

Otrzymano: 15 paź 2018

Przyjęty: 08 sty 2019

DOI: https://doi.org/10.2478/jdis-2019-0005

Słowa kluczowe
Sentiment analysis, Japanese reviews, Rule-based methods, Statistics-based methods, Tourism reviews

© 2019 Chuanming Yu, Xingyu Zhu, Bolin Feng, Lin Cai, Lu An, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Figure 1

Figure 2

Figure 3

The results of rule-based methods.

The classification results of all emotional categories.

The classification results of different methods.

The results of discarding words.

Sentiment Analysis of Japanese Tourism Online Reviews

Article Category: Research Paper

Data publikacji: 21 lut 2019

Zakres stron: 89 - 113

Otrzymano: 15 paź 2018

Przyjęty: 08 sty 2019

DOI: https://doi.org/10.2478/jdis-2019-0005

Słowa kluczoweSentiment analysis, Japanese reviews, Rule-based methods, Statistics-based methods, Tourism reviews

© 2019 Chuanming Yu, Xingyu Zhu, Bolin Feng, Lin Cai, Lu An, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Figure 1

Figure 2

Figure 3

The results of rule-based methods.

The classification results of all emotional categories.

The classification results of different methods.

The results of discarding words.

Słowa kluczowe
Sentiment analysis, Japanese reviews, Rule-based methods, Statistics-based methods, Tourism reviews