Future of Pervasive Health Workshop

Research Article

Towards a Federated Repository of Mobile Sensing Datasets for Pervasive Healthcare

  • @INPROCEEDINGS{10.4108/eai.16-5-2016.2263872,
        author={Jesus Favela and Luis A. Castro and Layla Michan},
        title={Towards a Federated Repository of Mobile Sensing Datasets for Pervasive Healthcare},
        proceedings={Future of Pervasive Health Workshop},
        publisher={ACM},
        proceedings_a={FUTURE OF PERVASIVE HEALTH WORKSHOP},
        year={2016},
        month={6},
        keywords={mobile sensing; dataset repository; data curation},
        doi={10.4108/eai.16-5-2016.2263872}
    }
    
  • Jesus Favela
    Luis A. Castro
    Layla Michan
    Year: 2016
    Towards a Federated Repository of Mobile Sensing Datasets for Pervasive Healthcare
    FUTURE OF PERVASIVE HEALTH WORKSHOP
    EAI
    DOI: 10.4108/eai.16-5-2016.2263872
Jesus Favela1,*, Luis A. Castro2, Layla Michan1
  • 1: CICESE
  • 2: Sonora Institute of Technology
*Contact email: favela@cicese.mx

Abstract

Mobile sensing is becoming a popular approach to infer patterns of activities and behavior to determine how they affect health and wellbeing. This data-driven approach to discovery has the potential to become a major tool in the field of epidemiology, aimed at determining the causes of disease in populations. These studies have motivated the creation of datasets with information opportunistically gathered from sensors in mobile devices. The nature of this data gathering effort raises a number of issues, such as the heterogeneity of the devices and sensors used, which hamper information sharing and integration needed to conduct longitudinal studies and validate and construct over previous results as new data becomes available and algorithms are improved. This paper proposes the development of an open access federated repository of datasets for preservation and sharing. We propose a process that involves data curation and integration into a unified schema from which researchers can query and use diverse dataset for comprehensive studies which could, for instance, compare two populations sensed in different periods and with somewhat different conditions.