TY - GEN
T1 - Sentiment Analysis of Standardization using Deep Belief Network
T2 - 4th Asia Pacific Conference on Research in Industrial and Systems Engineering: Building Business Resilience to Face the Challenge in Pandemic Era, APCORISE 2021
AU - Hartanto, Aries Agus Budi
AU - Zulkarnain,
AU - Surjandari, Isti
N1 - Publisher Copyright:
© 2021 ACM.
PY - 2021/5/25
Y1 - 2021/5/25
N2 - Free trade era requires increasing the competitiveness of local products in the global market, through standardization. The standardization policy is including how to plan, formulate, establish, implement, enforce, maintain, and supervise National Standard, e.g Indonesian National Standard called SNI. SNI is useful in order to create competitiveness and consumer protection. The consistency of standardization shows through standardization activity, that requires time and high resources. The number of SNI and the breadth of products distribution cannot be monitor simultaneously in the same years, also another obstacle in standardization activities. Therefore the aim of this study is to find a classification of standardization activity, which to becomes an important part of evaluation policy. The development of media plays a role in policy making, information and opinions from the media can change standardization's policy strategies. The contribution of this research is using text mining from standardization publication in media, to find useful knowledge. It's useful to build an input alternative, in the form of media sentiment analysis in standardization activity, that has never been done before. It gives an agile method for dealing with rapid changes in the standardization process. This study uses a deep belief network (DBN) method for the classification of media sentiment. Besides using DBN, this study also compares DBN with other classification methods, namely Naive Bayes (NB) and Support Vector Machine (SVM). These research results show the accuracy of the classification model with DBN reaches 77%, NB reaches 74% and SVM reaches 77%. Moreover, the results show that the most negative sentiment is 19% and the most positive sentiment is 29.20%. Both of the sentiments are the member of class about implementation and the mandatory regulation of SNI, and those aspects becoming media concentration. Standardization situation is expected to be captured as the output of this study so that it can contribute to improving the standardization policy in Indonesia.
AB - Free trade era requires increasing the competitiveness of local products in the global market, through standardization. The standardization policy is including how to plan, formulate, establish, implement, enforce, maintain, and supervise National Standard, e.g Indonesian National Standard called SNI. SNI is useful in order to create competitiveness and consumer protection. The consistency of standardization shows through standardization activity, that requires time and high resources. The number of SNI and the breadth of products distribution cannot be monitor simultaneously in the same years, also another obstacle in standardization activities. Therefore the aim of this study is to find a classification of standardization activity, which to becomes an important part of evaluation policy. The development of media plays a role in policy making, information and opinions from the media can change standardization's policy strategies. The contribution of this research is using text mining from standardization publication in media, to find useful knowledge. It's useful to build an input alternative, in the form of media sentiment analysis in standardization activity, that has never been done before. It gives an agile method for dealing with rapid changes in the standardization process. This study uses a deep belief network (DBN) method for the classification of media sentiment. Besides using DBN, this study also compares DBN with other classification methods, namely Naive Bayes (NB) and Support Vector Machine (SVM). These research results show the accuracy of the classification model with DBN reaches 77%, NB reaches 74% and SVM reaches 77%. Moreover, the results show that the most negative sentiment is 19% and the most positive sentiment is 29.20%. Both of the sentiments are the member of class about implementation and the mandatory regulation of SNI, and those aspects becoming media concentration. Standardization situation is expected to be captured as the output of this study so that it can contribute to improving the standardization policy in Indonesia.
KW - Deep belief network
KW - Sentiment analysis
KW - Standardization
KW - Text mining
UR - http://www.scopus.com/inward/record.url?scp=85143888950&partnerID=8YFLogxK
U2 - 10.1145/3468013.3468667
DO - 10.1145/3468013.3468667
M3 - Conference contribution
AN - SCOPUS:85143888950
T3 - ACM International Conference Proceeding Series
SP - 621
EP - 626
BT - Proceedings of the 4th Asia Pacific Conference on Research in Industrial and Systems Engineering
PB - Association for Computing Machinery
Y2 - 25 May 2021
ER -