Guiding Assistive-Technology Adaptations Through Intelligent Stream Mining of Patient Data

Guiding Assistive-Technology Adaptations Through Intelligent Stream Mining of Patient Data

William N. Robinson (Georgia State University, USA), Tianjie Deng (University of Denver, USA) and Andrea Aria (Georgia State University, USA)
Copyright: © 2020 |Pages: 40
DOI: 10.4018/978-1-7998-2310-0.ch011
OnDemand PDF Download:
No Current Special Offers


Users with cognitive impairments use assistive technology as part of a treatment plan. As the AT interface is manipulated, data stream mining techniques are used to monitor user goals. In this context, data mining aids caregivers in tracking user behaviors as they attempt to achieve their goals. Divergences over consecutive stream-mined models identify potential changes in user goal attainment, as the user learns his or her personalized emailing system. When a data-mined model diverges significantly from recent models, the user's behavior is flagged as a significant behavioral change. The specific changes in behavior are then characterized by analyzing model divergence as well as the underlying data. This chapter describes how divergence analysis of decision-tree and hidden Markov models can aid recognition and diagnoses of behavioral changes in support of AT adaptation, in a case study of cognitive AT for emailing. The technique may be more widely applicable to other behavior monitoring contexts.
Chapter Preview


One in 4 U.S. adults (61 million Americans) have a disability that impacts major life activities (Okoro et al., 2017). More than one million adults in the U.S. are diagnosed each year with cognitive impairments (CI) due to neurological disease or trauma. Currently, there are between 13.3 to 16.1 million Americans living with chronic brain disorders and associated CI (Alliance, 2001). In addition, approximately 4 million Americans have developmental disabilities that impact cognitive functioning (Services, 2002). Cognitive impairments prevent this large and growing segment of our society from fully integrating into society (McColl et al., 1998).

Assistive technology (AT) should help. However, studies have found that AT systems are abandoned by CI users at shockingly high rates (de Joode et al., 2010a; LoPresti et al., 2004; Wilson et al., 2001a; Wright et al., 2001). One major cause of abandonment is an eventual misalignment between: (1) user goals and abilities, and (2) the functionality delivered by the system. Research described in this chapter supports the monitoring of this relationship between user goals and their satisfaction by the system. In a case study, users are given an email system to aid in their activities of daily living (ADLs). At appropriate times, the system is adapted to meet the changing needs of the user. By monitoring a user’s event stream, post-clinic care teams and researchers can notice changes in user behavior that indicate that the system should be adapted. This work has mitigated this misalignment for dozens of clients engaged in cognitive rehabilitation. This technology is applicable to personalized user monitoring and generalizes to monitoring voluminous streams of event data.

Complete Chapter List

Search this Book: