Convex Optimization
Least-modules estimates for spatial autoregression coefficients
Journal of Computer and Systems Sciences International
Hi-index | 0.00 |
The least-modules method allows one to process efficiently the data with anomalously great errors. The paper was devoted to the Weiszfeld algorithm enabling one to solve approximately the variational problem corresponding to the least-modules method. Estimates of nonoptimality of the algorithms's iterations enabling one to estimate the quality of the approximate solution with anomalously great errors were obtained on the basis of the duality theory of the convex variational problems.