Issue No. 05 - May (1971 vol. 20)
C.C. Babu , IEEE
An algorithm for pattern classification using eigenvectors is presented. This algorithm provides an alternative viewpoint for obtaining a solution weight vector of certain pattern classification problems. Convergence of the algorithm is proved and the termination condition of iteration is given. An interesting point is that this algorithm does not require the inversion of matrices, instead it requires the computation of eigenvalues and eigenvectors of the matrix.
Eigenvectors, linear inequalities, linear inseparability, linear separability, pattern classification.
C. Babu and n. Wah-Chun Chan, "An Algorithm for Pattern Classification Using Eigenvectors," in IEEE Transactions on Computers, vol. 20, no. , pp. 575-578, 1971.