Investigating the Use of Semantic-Based Websites to Improve Recommendation Quality

In this paper, we aim to investigate the use of semantic-based websites to improve recommendation quality by testing a knowledge based recommendation system whose results completely depends on the product descriptions, on two different databases. In one of our relational MySQL databases, product descriptions are stored in form of RDF files and in the other one the data is stored in human language. We show that, the RS results are more accurate and intelligent when it is working with the semantic based database that stores product information in form of RDF graphs. Since, the welldefined data can help the recommendation system to analyze an extract the data better and make "smart" decisions.