156
Internet of Things (IoT)
Modeling can be classified based on the data used to build the model, type of model built,
and algorithms used for nature of model building (Cook et al. 2009).
Activity prediction and recognition—different type of sensors can recognize different
activities. It is based on the user requirement and the type of sensors used in the application for recognizing various activities. Figure 8.14 recognizes falling of a person. Table 8.1
shows the different sensors and the activity that it can recognize.
Activity model—different machine learning algorithms are applied to understand the
activity and to adapt to the changes based on the behavioral patterns. The pattern recognition can be in batch processing or steam data or can be for conditional patterns. Probabilistic
models, neural network, and fuzzy models are some of the techniques applied for activity
recognition.
FIGURE 8.14
Activity recognition—detecting falling of a person.
Reason
Act
Secure
Sense
HCI
Aml
FIGURE 8.13
Interaction of Aml system with other technologies.