GaitRec-Net: A Deep Neural Network for Gait Disorder Detection Using Ground Reaction Force

Table 9

State of the art of previous work.

Reference	Dataset	Methodology	No. of subjects	Classification & accuracy
[19]	Private dataset	Logistic regression; SVM & MARS	8	Binary class
				MARS = 88.3%; logistic regression = 68.5% & SVM = 84.8%
[20]	MFC data	SVM	58	83.3%
[18]	Private dataset	PCA + (SVM, KNN) & CNN	37	Binary class
				CNN = 91.9%; SVM = 67.6% & KNN = 48.7%
				Multiclass
				CNN = 83.8%; SVM = 51.4% & KNN =32.4%
[21]	Private dataset	PCA + linear SVM; RBF SVM	440	Binary class
				Linear SVM = 90.8%; RBF SVM = 89.1%
				Multiclass
				Linear SVM = 54.3; RBF SVM = 51.2%
[22]	Private dataset	KPCA + (SVM; ANN; random forest[RF])	239	Multiclass
				SVM = 89%; ANN = 90% & RF = 73%
Proposed method	GaitRec dataset	SVM; KNN; Naïve Bayes; 1D CNN	2295	Binary class
				SVM = 89.998%; KNN = 91.296%; Naive Bayes = 55.244% & 1D CNN = 91.624%