JSIAM Letters
Online ISSN : 1883-0617
Print ISSN : 1883-0609
ISSN-L : 1883-0617
Articles
Predicting the convergence of BiCG method from grayscale matrix images
Ryo OtaHidehiko Hasegawa
Author information
JOURNAL FREE ACCESS

2020 Volume 12 Pages 45-48

Details
Abstract

The convergence of the BiConjugate Gradient (BiCG) method depends on its input matrices. We tried to predict the convergence of BiCG method by applying a Convolutional Neural Network to matrices that had been converted to grayscale images. Using 875 real non-symmetric matrices in the SuiteSparse Matrix Collection, we applied the 5-fold cross-validation method and were able to predict convergence with an average accuracy that exceeded 80\% for all cases in the test collection.

Content from these authors
© 2015, The Japan Society for Industrial and Applied Mathematics
Previous article Next article
feedback
Top