Skip to content

Latest commit

 

History

History
23 lines (13 loc) · 1.26 KB

README.md

File metadata and controls

23 lines (13 loc) · 1.26 KB

Convex Structured Total Least Squares

This repository contains code for solving structured total least squares problems, i.e., problems for which both dependent and independent variables contain variable amounts of noise. The method is described in Malioutov & Slavov (2014).

comparing the performance of different methods for structured total least squares problem

Reproducing the analysis reported by Malioutov & Slavov (2014)

  • The figure comparing the performance of different methods can be generated by running the script compare_convex_STLS.m

Functions

  • The main function is struct_TLS_SDP_Aonly.m
  • The function RWNN_tls.m is a wrapper around the main function

Dependencies

  • The convex structured least squares algorithm uses several external packages listed below that can be downloaded separately. The main function struct_TLS_SDP_Aonly.m assumes that these packages are in the same directory and adds them to the path. The location of these packages can be changed and the change indicated in the subfunction addpaths in the body of the main function struct_TLS_SDP_Aonly.m
    • Yalmip
    • SDPT3-4.0
    • slra-0.5