Customer Classification Recommender to Support Personalised Service Level Agreements Across the Internet of Things

Abstract

While the Internet of Things (IoT) continues to be deployed in exciting application domains, the limited ability of customers to engage with novel technologies restricts the extent to which they are utilized. One significant factor influencing a customer's ability, and possibly their subsequent desire, to engage is the technical ability required to access such services. In our prior work, we have characterized online Internet consumers using customer classification guides which are not specific to the technical world; each guide characterizes customers according to categories influenced primarily by the type of home within which they live. The goal of the guides is to pre-empt purchasing behaviors. However, when considering the purchase of a Service Level Agreement (SLA) from an Internet Service Provider (ISP), we argue that considering this behavior on the basis of a household location and household characteristic may limit the overall suitability of the SLA which has been proposed - household savings, for example, may be high while technical ability low, leading to a mismatch in the service a customer needs. We therefore seek to remove household type as an influential attribute from the online customer characterization process, and instead focus on three customer qualities which are specific to the technologies used to access the Internet and a customer's technical competence. The overall contribution of this work is a model which can classify Internet users such that services can be offered without significant resource over- or under-provisioning, and the associated financial costs which such scenarios bring.