A	Time series containing information about observed human activities
$A^{\prime }$	Part of time series A used to train the recognition model M
$A″$	Part of time series A used to train the classification models $M_i$
$a_t$	Actual human activity at time step t
${\widehat{a}}_t$	Human activity determined at time step t by recognition algorithm
C	Classification algorithm (binary classifier)
$d_{i,t}$	Binary decision related to data transmission from sensor node i at time step t
D	Time series of of binary decisions
i	Identifier of sensor node
I	Set of identifiers of sensor nodes
$i^{*}$	Identifier of cluster head
M	Recognition model used by cluster head to recognize human activities
$M_i$	Classification model used by sensor node i for selecting the data that have to be transmitted
m	Number of time steps
n	Total number of sensor nodes
R	Activity recognition algorithm
S	Multivariate time series of preprocessed sensor readings from n sensor nodes
$S^{\prime }$	Part of time series S used to train the recognition model M
$S″$	Part of time series S used to train the classification models $M_i$
$S_t$	Set of data transmitted to cluster head from sensor nodes at time step t
$S_{i,t}$	Set of preprocessed sensor readings collected by sensor node i at time step t
$SM\left(Z\right)$	Sub-model for recognition of human activity based on data from sensor nodes belonging to subset Z
t	Time step
T	Time period (ordered set of time steps)