A Framework for Top Service Subscription Recommendations for Service Assemblers

Abstract

It is common practice today for small and medium business houses to assemble and host services, than hosting everything themselves. To cater to diverse market needs, these houses often need to subscribe to different services from different information providers. The service contracts and the range of features and facilities supported and provided by the providers vary widely. A non-trivial challenge for a service assembler is in deciding the set of information providers to subscribe to, given the heterogeneity in the offerings provided, the economics of the business model, the target set of customers in the market place and most importantly, the profit margin. We present in this paper, an automated framework that addresses this challenge and aids a service assembler with a cost-feature-performance balanced recommendation of the providers that can best serve his needs. The problem gets exacerbated since there can be multiple dimensions/categories of services (e.g., hotel, flight, and local conveyance in the travel domain) and there can be multiple relevant recommendations which may be of use for the service assemblers. We examine the service subscription recommendation problem from different perspectives and present algorithms for service assembly. Experimental results on small-scale real data as well as large-scale simulation data show the efficacy of our proposal.