Research on a Machine Learning-Based Method for Assessing the Safety State of Historic Buildings

Table 2

KCCA-SVM recognition accuracy of different kernel functions.

Datasets	Kernel functions	RMSE
		Minimum value	Average value	Maximum value
ORL	Gaussian kernel	8.322e − 2	8.581e − 2	8.676e − 2
	Sigmoid nucleus	8.693e − 2	8.825e − 2	8.991e − 2
	Polynomial kernel	9.592e − 2	9.674e − 2	1.201e − 1
Yale	Gaussian kernel	1.044e − 1	1.124e − 1	1.265e − 1
	Sigmoid nucleus	9.817e − 2	9.924e − 2	1.063e − 1
	Polynomial kernel	1.249e − 1	1.535e − 1	1.922e − 1
AR	Gaussian kernel	8.465e − 2	8.573e − 2	8.834e − 2
	Sigmoid nucleus	7.835e − 2	7.902e − 2	8.064e − 2
	Polynomial kernel	9.437e − 2	9.611e − 2	9.947e − 2
PIE	Gaussian kernel	7.916e − 2	8.111e − 2	8.259e − 2
	Sigmoid nucleus	8.146e − 2	8.297e − 2	8.481e − 2
	Polynomial kernel	1.098e − 1	1.244e − 1	1.367e − 1