Table 1.

Comparison of various ML methods adapted for neurodegenerative disorders such as Huntington or Parkinson disease to solve prediction and classification problems.

Work ref.	ML method	Learning approach	ML problem	Size of data set	Number of test subjects	Target group	Involve HD patients
[8]	ANN, MLP	Supervised	Classification	—	21	PD, healthy	No
[9]	RBFNN	Supervised	Regression	—	—	PD	No
[10]	DNN	Supervised	Classification	—	12	PD (8), healthy (4)	No
[11]	Decision tree, ID3	Supervised	Classification	195	31	PD (23), healthy (8)	No
[15]	Adaptive neurofuzzy	Hybrid		100	—	PD	No
[16]	Neurofuzzy system	Hybrid		—	—	ALS	No
[17]	Fusion of classifiers (Bayesian, SVM, k-nearest neighbor)	Hybrid		640	ALS (13), PD (15), HD (16), healthy (16)	ALS, PD, HD	Yes
[18]	Neurofuzzy system	Hybrid		—	—	Only survey was done	No
[19]	ANN + MLP, RBFNN	Hybrid		—	—	PD	No
[20]	Neurofuzzy system	Hybrid		—	—	—	No
[21]	PBL-McRBFN	Supervised	Classification	22,283	72 (50 PD, 22 healthy)	PD, healthy	No
[22]	Multistate Markov model	Hybrid		2500	72 (82 PD, 62 healthy)	PD, healthy	No
[23]	Random tree, (C-RT), ID3, binary logistic regression, k-NN, (PLS), (SVM)	Supervised	Classification	195	31 (23 PD, 8 healthy)	PD, healthy	No
[24]	FCM	Unsupervised	Clustering	195	—	PD	No