A comparative study of least squares, partial least squares, and ridge regression methods on transit time flowmeter measurement modeling
The regression method is an important factor in influencing the fitting performance of the TTUMF measurement model. Ordinary least squares (OLSR), partial least squares (PLSR), and ridge regression (RR) methods are compared in this paper. The polynomial modeling results indicate that the OLSR-derived model displays unsatisfactory model stability due to the collinearity among regressor variables, and the PLSR-derived model shows the best model stability.
Authors
- Jun Long
- Qingyuan Liu (ORCID: https://orcid.org/0000-0003-0061-2747)
- Chunyang Zhao (ORCID: https://orcid.org/0009-0004-9348-1735)
- Wei Guan (ORCID: https://orcid.org/0000-0002-6000-0489)
- Ping Wang (ORCID: https://orcid.org/0000-0002-2344-2544)
- Jintao Liu (ORCID: https://orcid.org/0000-0003-2215-2649)
- Xudong Wang (ORCID: https://orcid.org/0000-0002-9201-1692)
- Zhou Shi (ORCID: https://orcid.org/0000-0003-3914-5402)
Publication Details
- Published
- 2025-10-02
- DOI
- https://doi.org/10.1117/12.3083660
- Primary Topic
- Flow Measurement and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00