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

Multi-objective genetic algorithm-based sample selection for partial least squares model building with applications

Hideyuki Shinzawa1, Boyan Li, Takehiro Nakagawa

  • 1Department of Chemistry and Research Center for Near Infrared Spectroscopy, School of Science and Technology, Kwansei-Gakuin University, Hyogo 669-1337, Japan.

Applied Spectroscopy
|July 1, 2006
PubMed
Summary

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

Electric-Field-Driven Ferredoxin 1-Independent Cuproptosis Induction Overcomes Therapy-Induced Resistance in Glioblastoma.

ACS nano·2026
Same author

EXPRESS: Two-Dimensional Correlation Spectroscopy (2D-COS) Analysis of Near-Infrared (NIR) Spectra of Hydrogen-Bonded Elastomer Under Mechanical Stress.

Applied spectroscopy·2026
Same author

Process-Structure-Property Relationships in Meniscus-Confined Electrodeposition of 3D Copper Microcomponents.

ACS applied materials & interfaces·2026
Same author

Bif-1 upregulates autophagy to improve nitrite tolerance of Litopenaeus vannamei by interacts with Prohibitin-2.

International journal of biological macromolecules·2026
Same author

Rebinding drives cellular retention of small-molecule drugs with abundant targets.

Proceedings of the National Academy of Sciences of the United States of America·2026
Same author

Determination of the Blend Ratio of Styrene-Butadiene-Styrene (SBS)/Polyethylene (PE) and Polypropylene (PP)/PP Polymers Using a Portable Near-Infrared (NIR) Spectrometer Combined with the Classical Least Squares (CLS) Method.

Applied spectroscopy·2026

Multi-objective genetic algorithms (GAs) enhance partial least squares (PLS) models by removing erroneous samples. This approach improves model accuracy, robustness, and the detection of systematic errors in data analysis.

Area of Science:

  • Chemometrics
  • Data Science
  • Machine Learning

Background:

  • Partial Least Squares (PLS) is a widely used regression method.
  • Model performance can be degraded by systematic errors and outliers in training data.
  • Overfitting reduces the generalizability and robustness of predictive models.

Purpose of the Study:

  • To introduce multi-objective genetic algorithms (GAs) for optimizing sample selection in PLS model building.
  • To improve the performance, accuracy, and robustness of PLS models.
  • To enhance the detection capabilities for samples containing systematic errors.

Main Methods:

  • Application of multi-objective genetic algorithms (GAs) to identify and remove samples with systematic errors.
  • Utilizing training and validation sets to mitigate overfitting effects.

Related Experiment Videos

  • Development of an index to visualize systematic error factors using PLS and Pareto-optimal solutions.
  • Testing the methodology on near-infrared (NIR) spectral data.

    • Main Results:

    • Multi-objective GA significantly improved the performance of PLS models.
    • The GA-based sample selection enhanced the identification of samples with systematic errors.
    • Reduced overfitting led to more accurate and robust PLS models.
    • The introduced index effectively visualized factors contributing to systematic errors.

    Conclusions:

    • Multi-objective genetic algorithms offer a powerful approach to enhance PLS model building.
    • This method effectively addresses issues of systematic errors and overfitting in spectral data analysis.
    • The developed technique improves both model reliability and the ability to detect data anomalies.