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

Maximum likelihood principal components regression on wavelet-compressed data.

Marc N Leger1, Peter D Wentzell

  • 1Trace Analysis Research Centre, Department of Chemistry, Dalhousie University, Halifax, Nova Scotia B3H 4J3, Canada.

Applied Spectroscopy
|July 30, 2004
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

Augmented kurtosis-based projection pursuit: a novel, advanced machine learning approach for multi-omics data analysis and integration.

Nucleic acids research·2025
Same author

Assessing the precision of a detergent-assisted cartridge precipitation workflow for non-targeted quantitative proteomics.

Proteomics·2024
Same author

Analysis and Discrimination of Canadian Honey Using Quantitative NMR and Multivariate Statistical Methods.

Molecules (Basel, Switzerland)·2023
Same author

Coupling of multivariate curve resolution-alternating least squares and mechanistic hard models to investigate antibody purification from human plasma using ion exchange chromatography.

Journal of chromatography. A·2022
Same author

Beyond principal components: a critical comparison of factor analysis methods for subspace modelling in chemistry.

Analytical methods : advancing methods and applications·2021
Same author

Combinatorial projection pursuit analysis for exploring multivariate chemical data.

Analytica chimica acta·2021
Same journal

EXPRESS: Deterministic Compressed Sensing in Time-Domain Spectroscopy.

Applied spectroscopy·2026
Same journal

EXPRESS: Multi-Parameter Wavelength Characterization of Array Spectrometers Under Near-Limit Sampling Conditions.

Applied spectroscopy·2026
Same journal

EXPRESS: A Validated Reference Database for Twentieth-Century Cd-Based Pigments: Integrated Structural and Compositional Characterization.

Applied spectroscopy·2026
Same journal

EXPRESS: Two-Trace Two-Dimensional (2T2D-COS) in the Analysis of Brain Tissue Sample Preparation Method.

Applied spectroscopy·2026
Same journal

EXPRESS: Simplified Protocol for Analyzing Polarization Properties of Scanning Tunneling Microscope (STM) Light Emission Spectra at an Oblique Angle.

Applied spectroscopy·2026
Same journal

EXPRESS: Monitoring a Polyurethane Synthesis by Fiber-Coupled Attenuated Total Reflection Fourier Transform Infrared Spectroscopy and Multivariate Curve Resolution-Alternating Least Squares.

Applied spectroscopy·2026
See all related articles

Wavelet transforms (WT) significantly reduce computational demands for Maximum Likelihood Principal Component Regression (MLPCR) by compressing data. This method maintains predictive accuracy and offers substantial time savings, improving multivariate calibration models.

Area of Science:

  • Chemometrics
  • Data Science
  • Signal Processing

Background:

  • Maximum Likelihood Principal Component Regression (MLPCR) is an errors-in-variables method for multivariate calibration.
  • MLPCR's computational intensity, especially with large datasets and cross-validation, poses a significant challenge.
  • Measurement error is a critical factor in building accurate calibration models.

Purpose of the Study:

  • To introduce and evaluate wavelet transforms (WT) as an effective data compression technique for MLPCR.
  • To demonstrate how WT can mitigate the computational burden of MLPCR without compromising predictive performance.
  • To explore the relationship between spectral and error covariance matrices in both wavelet and spectral domains.

Main Methods:

  • Application of two-dimensional wavelet transforms (WT) for data compression prior to MLPCR analysis.

Related Experiment Videos

  • Utilizing the relationship between error covariance matrices in wavelet and spectral domains to account for WT effects.
  • Testing the WT-enhanced MLPCR approach on both simulated and experimental near-infrared (NIR) spectral data.

    • Main Results:

    • Significant data compression was achieved using WT, with improvements ranging from factors of 2 to 720.
    • The WT-compressed data in MLPCR resulted in reduced prediction errors compared to using raw data.
    • Favorable predictive ability was maintained despite substantial data compression.
    • Considerable time savings were realized in the MLPCR modeling process.

    Conclusions:

    • Wavelet transforms offer a powerful method for compressing data in MLPCR, addressing its computational limitations.
    • WT-based data compression enhances MLPCR efficiency and can lead to improved prediction accuracy.
    • This approach provides a practical solution for applying MLPCR to large datasets and complex error structures in chemometrics.