Application of Support Vector Machine in uncertainty Evaluation of geometric tolerance Measurement

Abstract

For some measured data with unknown distribution, when it is difficult to estimate the probability distribution and uncertainty of the measured results, based on a small amount of existing data, the probability density of the measured data is obtained based on the support vector machine method, and the standard uncertainty is calculated by random sampling with the obtained probability density distribution. The simulation results show that the "support vector" of Support Vector Machine（SVM） method is suitable for small sample data, and the accuracy of the optimal estimation and variance obtained on this basis is proved. Finally, taking the bearing roundness as the experimental object, the measurement uncertainty is calculated by the above method and compared with the results calculated by Monte Carlo method, which verifies the reliability and accuracy of the method.