Goal-oriented web-site navigation for on-line shoppers

Abstract

Web-sites for on-line shopping typically offer a vast number of product options and combinations thereof. While this is very useful, it often makes the navigation in the site and the identification of the "ideal" purchase (where the notion of ideal differs among users) a confusing, non-trivial experience. This demonstration presents ShopIT (ShoppIng assitanT), a system that assists on-line shoppers by suggesting the most effective navigation paths for their specified criteria and preferences. The suggestions are continually adapted to choices/decisions taken by the users while navigating. ShopITis based on a set of novel, adaptive, provably optimal algorithms for TOP-K query evaluation.

Navigating in complex mashed-up applications

“…However, we stress that this is the point where a semantic analysis, when can be performed efficiently, may take place. As our algorithm applies T opKP aths on a single application graph at a time, T opKP aths may implement an application-specific, semantic-aware analysis [4,5] that is incorporated in a transparent way within our algorithm.…”

“…information on pharmacies offering deals may be found in medical forums, payment may be done online, etc.). This implies that the number of possible relevant navigation flows in a MashAPP may be very large (even in a single application, the number of flows is large [4] and inter-connections between the applications further increase it). Some of these navigation options exploit the MashAPP structure in a much better way and are significantly better for the user than others (e.g.…”

“…Various works [4,6,5] considered analysis of navigation flows within the context of a single application. Most of these works analyze also the data manipulated by the WebApplication; as noted in [5,6], querying the combination of execution flows and data they manipulate, and the interplay thereof, incurs very high complexity.…”

“…Most of these works analyze also the data manipulated by the WebApplication; as noted in [5,6], querying the combination of execution flows and data they manipulate, and the interplay thereof, incurs very high complexity. The work in [4] took an approach similar to ours of querying application structure and suggested a PTIME evaluation algorithm for the restricted setting of a single isolated application. In contrast, we show that there is an inherent added hardness in the MashAPP settings, as if P = N P such PTIME algorithm does not exist here.…”

On models and query languages for probabilistic processes

“…Specifically, in [5,16,17] we have suggested a model and query language for Probabilistic Business Processes (BPs), which are used to depict the logical structure of business activities that are common e.g. in e-commerce and in Web Applications [5,19]. The model is an abstraction of the BPEL [8] standard and allows for an intuitive, graphical representation of the process.…”