A microservice-based architecture for an online product review analysis system / Mehdi Mohammadi

Since the emergence of Web 2.0 in late 1999, online product reviews have played a significant role in the e-commerce channels, social media, online forums, and online communities. The majority of studies in the area of product review analysis (PRA) systems have focused on evaluating and improving th...

全面介绍

Saved in:
书目详细资料
主要作者: Mehdi , Mohammadi
格式: Thesis
出版: 2021
主题:
在线阅读:http://studentsrepo.um.edu.my/13335/2/Mehdi_Mohammadi.pdf
http://studentsrepo.um.edu.my/13335/1/Mehdi_Mohammadi.pdf
http://studentsrepo.um.edu.my/13335/
标签: 添加标签
没有标签, 成为第一个标记此记录!
实物特征
总结:Since the emergence of Web 2.0 in late 1999, online product reviews have played a significant role in the e-commerce channels, social media, online forums, and online communities. The majority of studies in the area of product review analysis (PRA) systems have focused on evaluating and improving the techniques for review analysis to increase efficiency, accuracy, effectiveness, usefulness, ranking, spam detection, and feature extraction. The architecture of PRA systems is not the primary concern of research in this field. In this study, some of the implicit and explicit architectures of scientifically reported PRA systems were extracted to examine how the architectures have been evolving. One of the recently employed architectures in this area is the microservice architecture, aimed at achieving high flexibility and maintainability by developing loosely-coupled microservices. Current PRA systems do not fully utilise microservice architecture to achieve low-coupling in design. This research focuses on evaluating existing PRA architectures and proposing improvements on a microservice-based PRA architecture (Viscovery) in terms of reducing coupling. To evaluate the proposed architecture, two different methods, one of them is a quality model (MM4S model) and the other one involving descriptive techniques, were adopted. A prototype PRA system was developed based on the proposed architecture. The prototype system can retrieve reviews from multiple e-commerce websites and analyse the review text. The evaluation shows that the prototype system achieves lower coupling, as compared to a benchmarked microservice-based PRA system.