Mismatch robustness of linear prediction and its relationship to coding

Yasuo Matsuyama*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)


Source mismatch problems of linear least square predictors are discussed in terms of spectral distance measures and process distance measures. It is shown that the linear predictors are robust for several stability measures. The relationship of the predictor mismatch to linear predictive coding is also discussed, and a simpler alternative to Chaffee's rate distortion coder is given. In the Appendix, topological implications of distance measures are studied in order to supplement the robustness theorems in the text.

Original languageEnglish
Pages (from-to)237-262
Number of pages26
JournalInformation and control
Issue number3
Publication statusPublished - 1980
Externally publishedYes

ASJC Scopus subject areas

  • Engineering(all)


Dive into the research topics of 'Mismatch robustness of linear prediction and its relationship to coding'. Together they form a unique fingerprint.

Cite this