Aiming at the problem of low efficiency in the high-density measurement of electrical thickness of radome

an electrical thickness prediction method based on support vector machine regression model is proposed

which is characterized by accurate and efficient prediction. The data were extracted at different intervals to form 10 sets of training and testing sets for support vector machine regression model training

and the data show that the model constructed when the interval between the two points in the circumferential direction is 48 mm is accurate: the mean square error is 3.634 7

and the coefficient of determination is 0.972 83. Also validated by both multiple linear regression and BP neural network fitting

the results show that the data-driven support vector machine regression model has more accurate prediction results for small samples and nonlinear data

which verifies the feasibility of radome electrical thickness prediction

and the electrical thickness measurement points can be reduced by 87.5%.