Exploring web-based recommendation systems

Abstract

The presence of the web has been identified as the competitive playground for entrepreneurs to generate revenue; therefore many optimization techniques are pushing the professionals in the industry to find new ways how to acquire that competitive edge over their rivals. This study identities a set of techniques used for recommendation systems that can be adopted by web-based platforms, such as e-commerce, aiming to help sellers increase sales and buyers identify the best items that suits their individual needs, in an optimized trading environment. Some today define recommendation systems as "a serious business tool" (Schafer, Konstan, & Riedl, 2001). The field of recommendation systems is now used by more e-commerce platforms with the objective of optimizing their online business in areas such as user experience. In this project, recommendation systems will be investigated thoroughly. This will be done by investigating the different types of recommendation algorithms and success measurements. From this study, the necessary knowledge will be obtained to develop and implement these recommendation systems on a web platform and compare their performances in regards to the recommended items provided. These studies will serve as a base to draw conclusions on what type of recommendation systems online entrepreneurs should adopt in their business models, according to the circumstances they are dealing with. Different types (Leavitt, 2006) and techniques (Hamid Rastegari, 2008) of recommendation systems, such as content-based, knowledge-based, hybrid and collaborative filtering, are studied to observe the mechanisms of such systems; evaluating methods, applications and challenges (Hamid Rastegari, 2008) and lessons learned (RON KOHAVI, 2004) but also investigate whether "recommendation technology will boost ecommerce" (Leavitt, 2006) or not. Evaluation techniques are studied in detail (Dietmar Jannach, 2011) to better understand how items are ranked and chosen to be recommended to a specific customer.