Exploiting domain knowledge and public linked data to extract opinions from reviews
Rowida Alfrjani, Taha Osman, Georgina Cosma · 2017
Online opinions play an important role in supporting consumers make decisions about purchasing products or services. In addition, customer reviews allow companies to understand the strengths and limitations of their products and services, which aids in improving their marketing campaigns. Such valuable information can only be obtained via appropriate analysis of the opinions provided by customers who express their satisfaction with the products and shopping experience in the form of on-line textual reviews. This paper presents an approach to extract opinions from online reviews. The proposed approach constructs a knowledgebase of the problem domain, which is bootstrapped with relevant information from public Linked Datasets. The obtained results demonstrate that given a domain knowledgebase that is enriched with ground facts from public Linked Open Data with a combination with Natural Language process engine, the process of opinion extraction from unstructured content can be improved.