Table 2.

A list of common methodologies for data processing.

Methodology	Features	Task	Sensors	Test Set	Ref.
SVM and DT	Features extracted from filtered acceleration data samples: amplitude, time, statistics, orientation	Fall Detection	Wearable sensor: Tri-axial Accelerometer at waist	6 young adults and 2 elders performing 19 daily activities and 15 fall activities	[37]
kNN, NB, SVM and ANN	Vector magnitude of acceleration and angular velocity	Fall Detection	Wearable sensors: accelerometer, gyroscope and magnetometer at wrist and chest	17 people performing several daily activities	[102]
LSTM, GRU, SVM and kNN	Time series of accelerometer data	Fall detection	Wearable sensor: tri-axial accelerometer	23 adults and 15 elders performing several daily activities and falls	[40,103]
ANN	Spatio-Temporal Features	Anomaly detection in daily activities	Wearable sensors (accelerometer and gyroscope) and Ambient sensors	2 subjects performing 9 daily activities	[109]
Time series machine learning techniques	Time series data	Behavioral trend generation and forecasting	Sensors in the objects and Ambient sensors	4 subjects performing 6 daily activities	[110]
CT-HSMM	Stream of typed and time-stamped events	High level activities recognition	Sensors in doors and household appliances	7 activities, 28 days of observations	[111]
NB, SVM, RFs, DT, CNN, LSTM	Sensors data, activity, and context labels	Daily activity recognition	72 sensors: wearable sensors, object sensors and ambient sensors	4 subjects performing 7 daily activities	[106]
Multivariate Gaussian Distribution	Statistical features	Activity recognition	Ambient sensors: smart wall equipped with RFID sensors	4 subjects performing 12 real life daily activities	[70]
CNN	Time series	Abnormal behaviors detection	Wearable sensors	9 daily activities	[105]
RF, kNN	Spatial features	Daily activity recognition	Wearable sensor: accelerometer at chest	13 subjects performing 7 daily activities	[104]
CNN	Spatio-Temporal features	Daily activity recognition	Ambient sensors: depth camera	7 participants performing 21 sets of activities	[116]
NB, MLP, RF	Spatio-Temporal features	Daily activity recognition	Ambient sensors: RGB-D cameras	13 daily activities	[115]
ANN	Spatio-Temporal features	Daily activity recognition	Ambient sensors: depth cameras and acoustic sensors	17 subjects performing 24 daily activities	[114]
HMM	Spatio-Temporal features	Anomaly detection in daily activities	Wearable and ambient sensors	10 subjects performing daily activities over 3 months of observation	[112]
LSTM, RNN	Individual sensor events or group of sensor events in various time periods	Changes in behavioral patterns	IoT sensors: sensors in objects and furniture	6 elderly people observed at home over a period from 1.5 to 4 months	[108]
Unsupervised Learning	Spatio-Temporal features	Mild Cognitive Impairment Detection	Ambient sensors: motion sensor and door sensor	10 elderly people	[117]
Unsupervised Learning	Temporal features connected to temporal cluster of sensor events	Behavioral change detection	Ambient sensors: PIR sensors	Selection of data from Aruba data set: 28-day observation period	[118]