Pattern Recognition for Automated Healthcare Assessment Using Non-invasive, Ambient Sensors

In this paper, a solution for an automated healthcare assessment process is proposed. Non-invasive, ambient sensors are retrieving data from patients being in their home care treatment setups. The type of sensors is limited to the tracking of inertia, motion, and alcohol gas. Low-cost sensor prototypes are developed. They constantly measure the movement and the air around the patients. The Big Data generated in this way is used to retrieve patterns of activities. Different pattern recognition algorithms are tested and compared. The highest accuracy and reliability in assessing the data are support vector machines and feedforward neural networks with a performance of 90 % probability in identifying the correct patients’ activities over the test period. In this paper, the setup of the sensor prototypes, the data handling, and the data analytics are discussed.

[1]  Chih-Jen Lin,et al.  A comparison of methods for multiclass support vector machines , 2002, IEEE Trans. Neural Networks.

[2]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.

[3]  Diogo R. Ferreira,et al.  Preprocessing techniques for context recognition from accelerometer data , 2010, Personal and Ubiquitous Computing.

[4]  V. Garms-Homolova.  Assessment für die häusliche Versogung und Pflege : resident assessment instrument - home care (RAI HC 2.0) , 2002 .

[5]  Patrick Olivier,et al.  Feature Learning for Activity Recognition in Ubiquitous Computing , 2011, IJCAI.

[6]  Rolf Dornberger,et al.  Healthcare Assessment Questions Non-invasive Ambient Sensors , 2016, 2016 IEEE International Conference on Healthcare Informatics (ICHI).

[7]  P. Enright,et al.  The six-minute walk test. , 2003, Respiratory care.

[8]  T. Wagner,et al.  Self-Reported Utilization of Health Care Services: Improving Measurement and Accuracy , 2006, Medical care research and review : MCRR.

[9]  J. Byles,et al.  A thorough going over: evidence for health assessments for older persons , 2000, Australian and New Zealand journal of public health.

[10]  Paul Smolensky,et al.  Information processing in dynamical systems: foundations of harmony theory , 1986 .

[11]  Kristof Van Laerhoven,et al.  Combining wearable and environmental sensing into an unobtrusive tool for long-term sleep studies , 2012, IHI '12.

[12]  Mohammad Mehdi Homayounpour,et al.  A brief survey on deep belief networks and introducing a new object oriented MATLAB toolbox (DeeBNet) , 2014, ArXiv.

[13]  Remco R. Bouckaert.  Naive Bayes Classifiers That Perform Well with Continuous Variables , 2004, Australian Conference on Artificial Intelligence.

[14]  Maureen Schmitter-Edgecombe,et al.  Automated Cognitive Health Assessment Using Smart Home Monitoring of Complex Tasks , 2013, IEEE Transactions on Systems, Man, and Cybernetics: Systems.

[15]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[16]  G. Onder,et al.  Minimum Data Set for Home Care: A Valid Instrument to Assess Frail Older People Living in the Community , 2000, Medical care.

[17]  Sotiris B. Kotsiantis,et al.  Supervised Machine Learning: A Review of Classification Techniques , 2007, Informatica.

[18]  Rolf Dornberger,et al.  Sensor-Based Tracking and Big Data Processing of Patient Activities in Ambient Assisted Living , 2016, 2016 IEEE International Conference on Healthcare Informatics (ICHI).

[19]  Allan Pinkus,et al.  Multilayer Feedforward Networks with a Non-Polynomial Activation Function Can Approximate Any Function , 1991, Neural Networks.