What is the specific algorithm for finding eigenvalues and eigenvectors in matlab?

Eig(A) mainly uses QR algorithm. If a is symmetric, use symmetric QR algorithm (divide and conquer); If feature vectors are needed, they can be used).

Eig(A, b) uses QZ algorithm. If it is detected that A symmetry and B symmetry are positive definite, Coleski decomposition is performed on B, and then symmetric QR algorithm is used.

The algorithm of svd is similar to the symmetric QR algorithm.

These are not understandable in a few words. You have to take the course of matrix calculation (numerical linear algebra) to know the simplest algorithm decades ago. Many new technologies have been added to the algorithms in the software, but the basic methods are similar.