期刊名称：Journal of King Saud University @?C Computer and Information Sciences
印刷版ISSN：1319-1578
出版年度：2019
页码：1-9
DOI：10.1016/j.jksuci.2019.12.009
出版社：Elsevier
摘要：Banking institution is a critical part of any country’s economy, which provides a variety of services for an individual or a business entity. Due to vast availability of banks and their services, it is a cumbersome task for a person to choose a bank for his/her specific need. In this regard, we introduced a computational framework for ranking a set of banking institutions based on people’s opinion. The introduced framework is based on aspect-based sentiment analysis and multi-criteria decision making (MCDM) approaches. To evaluate our methodology, we developed a sentiment dataset comprising reviews on four Indian nationalized banks, whose quality is evaluated using 9 aspects/attributes. In order to come up with holistic ranking, we employed simple majority voting on ranking obtained using three multi-criteria decision making methods viz. analytic hierarchy process, VIKOR, and fuzzy multi-attribute decision making. We observed that final ranking obtained using our method has a strong resemblance with real-life outreach of selected four Indian banks.
关键词：Analytic hierarchy process ; Aspect;level sentiment analysis ; Information system ; Online reviews ; MCDM
