The number of e-commerce sites is growing at an astounding rate. Low personnel overhead, just-in-time supply, and the widespread acceptance of online credit card payments make large e-stores a viable business model. Indeed, the economics of small or no inventories seems to be a compelling force toward mammoth stores; successful stores such as Amazon have shifted from a focused line of products to selling quite diverse and heterogeneous items. E-commerce portals are one of the most active and important Internet application areas, yet selecting a product to buy is frequently a frustrating experience because of the size of inventory, and, most importantly, because customers do not know exactly the specific item they want, but are rather looking for the item that best fits their individual requirements. This situation requires system assistance in browsing and exploration as opposed to retrieval based on a precise specification, which is the paradigm of search tools (text retrieval and database queries) supplied by traditional technology and used in most portals.