Machine-Learning-Based Human Fall Detection Using Contact- and Noncontact-Based Sensors

Table 1

Parameters used for ML methods.

Architecture used	Parameters
SVM method	Kernel = “radial basis function”
	C = 1
	Gamma = 0.1
	Tolerance = 0.001
RF method	Estimators = 100
	Criterion = “Gini”
	Sample split = 2
	Min. sample leaf = 1
	Max. depth = none
MLP method	Activation = “ReLU”
	Solver = “Adam”
	Max. number of iterations = 200
	Tolerance = 0.0001
	Shuffle = true
	Initial learning rate = 0.001
	Beta1 (exponential rate decay) = 0.9
	Beta2 (exponential rate decay) = 0.999
	Epsilon (numerical stability measure) = 10−8