Semantically enhanced and efficient location services for preserving mobile consumer's privacy

In Mobile Commerce applications, such as location-based advertising, merchants use consumers' information to send them personalized advertisements. These applications provide convenience to consumers and competitive advantages to merchants. However, the improper use of consumers' information presents a serious threat to their privacy.; The problem environment for this research is the context of a Location Service (LS) that maintains the two types of information used for personalization: consumer profiles and location information.; First, we provide support for the choice of the LS to maintain this information and for the rationale for the consumer to entrust it with their information. Then, we address three problems in that environment: (1) How to build a system for enforcing mobile consumers' privacy preferences that supports granular representation and spatio-temporal constraints; (2) How to extend that system to support semantically enhanced preferences; and (3) How to improve the efficiency of data management at the LS.; We propose a solution for the first problem that includes an access control model for moving objects and consumer profiles and a mechanism that enforces spatio-temporal policies. The mechanism includes a new variation of the trie structure referred to as the Adaptive Search Multiway trie (ASM-trie): Our evaluation study shows that the ASM-trie has a positive impact on the efficiency of the enforcement mechanism. We then extend that work by employing a semantic representation and reasoning. We consider the domain of promotions as an illustrative example. We show how to model consumer preferences and merchant promotions in an ontology using a Semantic Web language and how to enforce these preferences by matching them with merchant queries using reasoning techniques.; In addressing the third problem, we present a unified indexing scheme that improves query performance by partitioning consumer information into clusters and mapping each cluster to a moving objects tree that represents the location information of the consumers in that cluster. The indexing scheme answers queries using a new classification approach with selectable accuracy.