Framework of uCRM based on Context-aware characteristics of U-commerce and personalization service using contextual data

Abstract

Ubiquitous Computing is quickening a new era of U-commerce where buyers and sellers will literally be able to commerce anytime, anywhere, and any way they like. The U-commerce service is “context-aware,” and it focuses more on actively sensing different customer’s roles through both time and location specificity. In U-commerce environment, we can make decisions proactively and intelligently by automatically detecting users’ contextual data such as time, identity, location, entity. Context-aware technology can provide personalization services that reference the user’s context and preferences. Proactive service and high personalization will enable a great number of improvements in the current CRM processes and open a new area of customer satisfaction. uCRM must pay due regard to ‘context-aware’ characteristics of U-commerce. This thesis consists of three studies - practical framework of uCRM, methodologies for contextual data, and personalization system of tangible product and intangible service. The first, we proposed the practical framework of uCRM based on context-aware characteristics of U-commerce. In the framework, all customer information is stored in context datawarehouse (contextDW). These tremendous data are analyzed with context mining containing data mining and web mining techniques. With analysis results, marketing strategies will be developed and delivered to proper customer segment. The second, because contextual data contains symbolic values such as location and entity, we proposed a hybrid learning methods of case-based reasoning and neural network, which we named CANSY, for symbolic features. CANSY includes a global feature weighting method by neural network for case-based reasoning and uses value different metric (VDM) for measuring the distances for nominal features. Finally, we present not only framework and methodologies but also experimental results and applications. We proposed personalization systems for tangible products and intangible s...