Authors: Alexander Felfernig
Addresses: Computer Science in Production, Operation and Management, Departments of Informatics, University Klagenfurt, Universitaetsstrasse 65-67, A-9020, Klagenfurt, Austria
Abstract: Complex product assortments offered by online stores and electronic marketplaces make the identification of appropriate solutions a challenging task. Customers can differ greatly in their level of product domain knowledge. Therefore, recommender applications that support the product selection process are required. Knowledge-based recommender technologies allow flexible mapping of product, marketing and sales knowledge to the formal representation of a knowledge base and thus support customer-oriented sales dialogues in online selling environments. Although these technologies are frequently applied in industrial settings, related development processes are still very time-consuming. This paper presents the Koba4MS environment that supports graphical knowledge acquisition for web-based recommender applications and thus improves the effectiveness of related development and maintenance processes. The results of an experiment that investigated the effects of applying automated testing and debugging techniques to recommender knowledge base development clearly show the applicability and effectiveness of our approach in terms of time savings and reduced error rates.
Keywords: web-based recommenders; recommender applications; web engineering; knowledge acquisition processes; testing; debugging; experimental evaluation; online selling.
International Journal of Web Engineering and Technology, 2007 Vol.3 No.3, pp.329 - 351
Available online: 16 Jan 2007 *Full-text access for editors Access for subscribers Purchase this article Comment on this article