[Retracted] Task-Oriented Intelligent Solution to Measure Parkinson’s Disease Tremor Severity

Table 3

Classifiers’ hyperparameters search spaces.

Classifier	Hyperparameters search spaces
ANN-MLP	batch_size: [32, 64, 512]
	Epochs: [200, 300]
	Neurons: Integer (60, 100)
	Optimizer: [SGD, RMSprop, Adam, Adadelta, Adagrad, Adamax, Nadam]
	Activation: [relu, tanh, selu, elu, exponential]
KNN	n_neighbors: Integer (1, 20)
	Weights: [Distance, uniform]
	Algorithm: [Brute, ball_tree, kd_tree]
	Metric: [Minkowski, euclidean, manhattan]
	leaf_size: Integer (1, 20)
	p: Integer(1, 2)
RF	n_estimators: Integer(10, 250)
	max_features: Integer(1, 102)
	max_depth: Integer(5, 100)
	min_samples_split: Integer(2, 20)
	min_samples_leaf: Integer(1, 20)
	Criterion: [gini, entropy]
DT	max_features: Integer(1, 102)
	max_depth: Integer(5, 100)
	min_samples_split: Integer(2, 20)
	min_samples_leaf: Integer(1, 20)
	Criterion: [gini, entropy]
LR	Penalty: [l2, none]
	C: [1e − 2, 1e − 1, 1e0, 1e1]
	Solver: [Newton-cg, lbfgs, sag, saga]
	max_iter: Integer(1, 1000)
SVM	C: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
	Gamma: [0.1, 0.01, 0.001]
	Degree: (1, 5)
	Kernel: [Linear, poly, rbf, sigmoid]