Extracting product unit attributes from product offers by using an ontology

Comparing products and prices of on-line offers is important to e-shoppers and on-line retailers. The e-shoppers compare products and prices for finding the best product or price whereas the retailers need to compare their own products and prices to those of their competitors in order to remain competitive. As there is a wide range of product offers available on the Web the product and price comparison need to be done by an automated approach. One challenging task of such an approach is the automated assignment of different product offers which refer to the same real-world product. Products are compared and assigned based on the calculation of the pairwise similarity of their attributes. Thus, the attributes need to be identified, extracted and harmonised. In the case of the comparison of unit attributes harmonisation means to convert the attributes and their values into the same unit. The contribution of this paper is an approach and a corresponding ontology for identifying, extracting and harmonising unit attributes of product offers. The unit attribute ontology is introduced, the extraction process is described and the adequacy of the approach is validated through an experiment.