ABSTRACT
ABSTRACT This chapter provides an overview of the machine learning (ML) concepts in the field of ambient assisted living (AAL). The ML techniques aim at structuring cognitive information stemming from raw data by means of a computer. Such data may be collected, for instance, by either patients or health care professionals (HCPs) during provided AAL services not only in the patient’s home but also in medical and occupational environments; these data include, for example, activities and medication reminders, objective measurement of physiological parameters, feedback based on observed patterns, questionnaires, and scores. Both patients and HCPs are sources of raw clinical data that require computational processes that give rise to useful information capable of supporting clinical decision making. This chapter describes ML in terms of learning concepts emphasizing the follow approaches: supervised, unsupervised, semisupervised, and reinforcement learning. In addition, the principles of concept classification are explained, and the mathematical concepts of several methodologies are presented, such as neural networks, support vector machine, and fuzzy logic, among other techniques. Finally, an approach based
CONTENTS
Introduction ................................................................................................................................. 536 Machine Learning ....................................................................................................................... 536
Supervised Learning .............................................................................................................. 537 Unsupervised Learning ......................................................................................................... 538 Semisupervised Learning ...................................................................................................... 538 Reinforcement Learning ........................................................................................................ 539
Classification ................................................................................................................................540 Artificial Neural Networks ................................................................................................... 541 Self-Organizing Map ..............................................................................................................544 Fuzzy Logic .............................................................................................................................545 Support Vector Machine ........................................................................................................546 k-Nearest Neighbor ................................................................................................................547 k-Means Clustering ................................................................................................................548 Linear Discriminant Analysis ...............................................................................................548 Principal Component Analysis ............................................................................................550 Independent Component Analysis ...................................................................................... 551
Approximator Fusion ................................................................................................................. 552 Conclusion ................................................................................................................................... 553 Acknowledgments ...................................................................................................................... 553 References ..................................................................................................................................... 553
on the fusion of several single methods is described for situations dealing with multiple learning models. The methods that are presented in this chapter were selected according to daily needs in medical data monitoring for the welfare of patients.
