Cloud-based platform for personalization in a wellness management ecosystem: Why, what, and how

Abstract

Offering personalized services through dynamically formed ecosystems is essential to personal well ness management. In this paper, we present the design of a cloud-enabled platform to facilitate the collection and delivery of evidence for personalization in a multi-provider ecosystem environment. In addition, the platform also provides essential building blocks of personalization services: smarter analytics for active personalization and dynamic provisioning. While the former common service takes charge of inferring user well ness risks from multiple data sources on the fly and making risk-driven recommendations, the latter common service determines optimal platform pricing and resource allocation given the constraint of acceptable quality of service.

Big Data computing and clouds: Trends and future directions

“…Users can access the models with Web browser technologies to compose their data mining solutions. Existing work also advocates the use of PMML as a language to exchange information about predictive models [105].…”

“…They discuss the technical challenges on isolating analytical artefacts. Hsueh et al [105] discuss issues related to pricing and Service Level Agreements (SLAs) on a platform for personalisation in a wellness management platform built atop a Cloud infrastructure. Krishna and Varma [25] envision two types of services for Cloud analytics: (i) Analytics as a Service (AaaS), where analytics is provided to clients on demand and they can pick the solutions required for their purposes; and (ii) Model as a Service (MaaS) where models are offered as building blocks for analytics solutions.…”

An integrated personalization framework for SaaS-based cloud services

“…Hsueh et al [11] proposed a cloud-enabled platform in order to facilitate the collection and delivery of evidence for personalization in a multi-provider ecosystem environment. This platform provides basic personalization services, smart analytics for active personalization and dynamic provision.…”

“…This noise is mixed into the user profile and usage data, and it is important to apply algorithms to clean the data or reduce the noise to an acceptable level. Hsueh et al [11] proposed an algorithm to filter statistical noise from collected user profile and behavior logs. Based on the PWR-wide risk stratification algorithm, and according to concept usage patterns, if one new usage data demonstrates 100% deviation from the pattern, we define it as a noise.…”

Next Generation Wellness: A Technology Model for Personalizing Healthcare

“…One approach we have taken is to create a data quality monitor to determine whether an identifi ed risk group is suffi ciently representative to be used for predicting risk [ 39 ]. Specifi cally, the monitor follows possible sources of prediction errors in three major categories: risk group noise, case ambiguity, and noise-adjusted case ambiguity.…”