AnalyticsServiceDesign - overview
This proposal on active characterization of personal wellness status and recommendation includes the components to handle the following challenges: First, the system needs to digest not only the patient’s personal wellness records (e.g., questionnaires and physio-social-psycho measurements), but also how these records are related to others (i.e., topological information in a risk inference network) and whether there exist any changes in the dynamic wellness system (i.e. change detection). Second, the computation burden incurred by the rapid accumulation of personal wellness records asks for techniques that can reasonably estimate wellness risks from a subset of observations. In addition, we propose a backend proactive learning mechanism to handle the reliability issues and establish a safety net for further analysis. These proposals altogether will facilitate the development of a personalization engine, which serves as the power center of all the different personalized services surrounding users and provides a benchmark service on the effectiveness of “adaptable programs”on personal wellness management. These can also serve as the foundation of solutions that can benefit from personalization, e.g., individualized medicine and wellness management, incentive scheme design, wellness status-sensitive localization services, social-aware enterprise health modeling.
The development of this engine entail a number of novel service designs. In particular, we focus on generating individualized guidelines, online active compliance feedback strategy, and proactive intervention plan adaptation.
Another important extension of personalization analytics service design research is to capture signals from social interactions during the follow-up period to improve compliance analysis at both the personal and case management level, enhance service delivery quality, and perform user-centric optimization for service recommendation and composition.