Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Boosting partial least squares.

M H Zhang1, Q S Xu, D L Massart

  • 1ChemoAC, Department of Pharmaceutical and Biomedical Analysis, Pharmaceutical Institute, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium.

Analytical Chemistry
|March 1, 2005
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

[The association between daily housework and all-cause mortality among rural older adults in Miyun District of Beijing].

Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi·2026
Same author

[Ovarian collision tumor with postmenopausal hyperandrogenism: a case report].

Zhonghua nei ke za zhi·2024
Same author

[A case of acute inhalation dinitrogen tetroxide poisoning].

Zhonghua lao dong wei sheng zhi ye bing za zhi = Zhonghua laodong weisheng zhiyebing zazhi = Chinese journal of industrial hygiene and occupational diseases·2024
Same author

[Primary cardiac synovial sarcoma: a clinicopathological analysis of five cases].

Zhonghua bing li xue za zhi = Chinese journal of pathology·2024
Same author

[Repair methods of complex facial defect wounds involving paranasal sinuses and their clinical effectiveness].

Zhonghua shao shang yu chuang mian xiu fu za zhi·2023
Same author

[Random Integration Analysis of Recombinant Adeno-Associated Virus 6 Packaged in Sf9 Insect Cells].

Molekuliarnaia biologiia·2023
Same journal

Heterojunction Gate-Empowered OPECT Aptasensing: A Valid Protocol for Realizing High Current Gain at Low Electron Donor Dependency.

Analytical chemistry·2026
Same journal

Development of a Tapping-Mode Scanning Probe Electrospray Ionization Platform for High-Sensitivity and Long-Term Stability in Single-Cell Mass Spectrometry Imaging of Tissue.

Analytical chemistry·2026
Same journal

A Solid-State Near-Infrared Fluorescent Probe by a Synergistic Extended Conjugated System for Detecting Cys with Long-Term Imaging in Orthotopic Bladder Cancer.

Analytical chemistry·2026
Same journal

Air-Based Porous Array Dielectric Barrier Discharge Ionization Source for Explosive Trace Detection.

Analytical chemistry·2026
Same journal

PSAQ<sup>+1</sup>: Absolute Protein Quantification Using a <sup>13</sup>C<sub>1</sub>-Labeled Protein Standard, Coisolation of Peptide Pairs and LC-PRM.

Analytical chemistry·2026
Same journal

Heterojunction-Enhanced Interfacial Evanescent-Tunable Fiber Optic Probe for Amplification-free CRISPR/Cas12a-Based Rapid and Ultrasensitive Detection of MPXV.

Analytical chemistry·2026
See all related articles

Boosting Partial Least Squares (BPLS) combats overfitting in multivariate calibration. This novel method combines shrunken PLS models iteratively, improving accuracy and resistance to overfitting compared to classical PLS.

Area of Science:

  • Chemometrics
  • Statistical Modeling
  • Machine Learning

Background:

  • Partial Least Squares (PLS) is widely used in multivariate calibration.
  • Overfitting is a significant challenge when applying PLS, leading to reduced model accuracy and generalizability.
  • Existing methods often require careful selection of the number of PLS components, which can be complex.

Purpose of the Study:

  • To introduce a novel method, Boosting PLS (BPLS), to address overfitting in multivariate calibration.
  • To develop an iterative approach that combines shrunken PLS models to enhance robustness.
  • To provide criteria for determining key BPLS parameters: shrinkage value and iteration number.

Main Methods:

  • BPLS iteratively constructs shrunken PLS models, each with a single component.

Related Experiment Videos

  • Subsequent models are built upon the residuals not explained by preceding models.
  • The method incorporates shrinkage to regularize individual PLS components.

    • Main Results:

    • BPLS demonstrated superior resistance to overfitting across seven real-world datasets.
    • The proposed method maintained accuracy comparable to classical PLS.
    • BPLS eliminates the need for selecting the optimal number of PLS components.

    Conclusions:

    • BPLS offers a robust alternative to classical PLS for multivariate calibration.
    • The iterative and shrunken component approach effectively mitigates overfitting.
    • BPLS provides a more reliable and accurate modeling strategy for complex datasets.