Research on Machine Learning Algorithm Based on Contour Matching Modal Matrix

Abstract

Abstract For the problem of contour matching modal matrix optimization, a combination of contour matching and gradient algorithm is proposed as an optimization search method. The basic idea of regression estimation and machine learning set is the same, both of them reconstruct the estimation of samples by estimating the machine learning clustering coefficients, and the main difference is the different models chosen. Although contour matching has the advantage of global optimization search, it has the defects of easy prematureness and poor local optimization search performance. Therefore, this paper combines it with the gradient algorithm, which not only speeds up the search in the gradient algorithm and ensures that the method converges to the global optimal solution, but also achieves the global convergence of the method and the high efficiency of the computational speed by keeping the optimal solution of the iterative process.