TY - GEN
T1 - Abstract generation system for Chinese articles and reviews of 3C products
AU - Chen, Yi Ting
AU - Wang, Tzone I.
N1 - Publisher Copyright:
© 2019 Association for Computing Machinery.
PY - 2019/10/18
Y1 - 2019/10/18
N2 - The Internet has become popular and convenient. Product articles and reviews are written by people on digital media platforms such as Facebook, PTT, Mobile01, and Apple Daily News. Most people read many articles and reviews on digital media when they want to buy a product. However, an overwhelming number of articles and reviews of products is available on the Internet, and a prospective buyer can become confused. People should organize these articles and reviews before deciding on whether to buy a product or not. Therefore, it is essential to summarize the available data and provide customers useful information.Many researchers have investigated this task, but most studies have been focused on English reviews. This work focuses on Chinese articles and reviews in digital media, and propose a system designed to summarize data from digital media. When a user is interested in a product, the system extracts features and opinion words of the product from review articles and uses these features to identify sentences highly related to the product. After obtaining these sentences, the approach in this work selects top 20 important sentences to form the summary of the product, which is presented to the user. This work conducts several experiments to compare the effectiveness of TextRank, Luhn’s method, and the proposed approach. Among them, the approach proposed in this work exhibits the best performance.
AB - The Internet has become popular and convenient. Product articles and reviews are written by people on digital media platforms such as Facebook, PTT, Mobile01, and Apple Daily News. Most people read many articles and reviews on digital media when they want to buy a product. However, an overwhelming number of articles and reviews of products is available on the Internet, and a prospective buyer can become confused. People should organize these articles and reviews before deciding on whether to buy a product or not. Therefore, it is essential to summarize the available data and provide customers useful information.Many researchers have investigated this task, but most studies have been focused on English reviews. This work focuses on Chinese articles and reviews in digital media, and propose a system designed to summarize data from digital media. When a user is interested in a product, the system extracts features and opinion words of the product from review articles and uses these features to identify sentences highly related to the product. After obtaining these sentences, the approach in this work selects top 20 important sentences to form the summary of the product, which is presented to the user. This work conducts several experiments to compare the effectiveness of TextRank, Luhn’s method, and the proposed approach. Among them, the approach proposed in this work exhibits the best performance.
UR - http://www.scopus.com/inward/record.url?scp=85077738756&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85077738756&partnerID=8YFLogxK
U2 - 10.1145/3366650.3366654
DO - 10.1145/3366650.3366654
M3 - Conference contribution
AN - SCOPUS:85077738756
T3 - ACM International Conference Proceeding Series
SP - 32
EP - 36
BT - ICCBD 2019 - 2019 the 2nd International Conference on Computing and Big Data, Workshop CSEA 2019
PB - Association for Computing Machinery
T2 - 2nd International Conference on Computing and Big Data, ICCBD 2019 and its Workshop the International Conference on Computer, Software Engineering and Applications, CSEA 2019
Y2 - 18 October 2019 through 20 October 2019
ER -