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

Development of a new regression analysis method using independent component analysis.

Hiromasa Kaneko1, Masamoto Arakawa, Kimito Funatsu

  • 1Department of Chemical System Engineering, The University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-8656, Japan.

Journal of Chemical Information and Modeling
|March 7, 2008
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

Machine learning-based prediction of polyvinyl alcohol product viscosity and design of optimal process conditions.

Analytical sciences : the international journal of the Japan Society for Analytical Chemistry·2026
Same author

Data-Driven Design of Organic Semiconductors Exhibiting Low Reorganization Energy via Hierarchical Variational Autoencoders, Gaussian Mixture Regression, and Bayesian Optimization.

Journal of chemical information and modeling·2026
Same author

Generation of Molecules Near the Applicability Domain Boundaries of Property Prediction Models.

Journal of chemical information and modeling·2026
Same author

A general framework for extrapolation-aware prediction reliability in forward and inverse analyses of Gaussian mixture regression models.

Analytical sciences : the international journal of the Japan Society for Analytical Chemistry·2026
Same author

Robust machine learning and ensemble learning approach to predict variation in experimental data for multiple measurements and anomalies.

Analytical sciences : the international journal of the Japan Society for Analytical Chemistry·2026
Same author

Machine Learning Models Predicting Solubility and Polymerizability of Polyimides Considering Multiple Monomers for CO<sub>2</sub> Separation Membranes.

Molecular informatics·2026
Same journal

tmGNN-XAI: An Explainable Graph Neural Network Tool for Predicting Electronic Properties of Transition Metal Complexes from SMILES.

Journal of chemical information and modeling·2026
Same journal

QSAR in the Browser: An Interactive Cheminformatics Web Application.

Journal of chemical information and modeling·2026
Same journal

FoldDoF: Utilizing the Primary Degrees of Freedom of Protein Backbone for Geometric Modeling and Generation.

Journal of chemical information and modeling·2026
Same journal

Derisking Affinity Optimization for Macrocycles and Cyclic Peptides: High-Precision Free Energy Simulations across Five Diverse Targets.

Journal of chemical information and modeling·2026
Same journal

An End-User Audit of Reproducibility, Data Leakage, and Overfitting of the Top-Ranked ADMET Prediction Models in TDC Leaderboards.

Journal of chemical information and modeling·2026
Same journal

PFASGroups: An Open-Source Framework for Automated Identification, Structural Classification, and Prioritization of Per- and Polyfluoroalkyl Substances.

Journal of chemical information and modeling·2026
See all related articles

Independent Component Analysis-Multiple Linear Regression (ICA-MLR) offers superior predictive accuracy for quantitative structure-property relationships. This novel method effectively extracts significant components from molecular descriptors for analyzing aqueous solubility.

Area of Science:

  • Computational Chemistry
  • Chemometrics
  • Machine Learning in Chemistry

Background:

  • Quantitative Structure-Property Relationship (QSPR) analysis is crucial for predicting chemical properties.
  • Traditional methods like Partial Least Squares (PLS) have limitations in extracting complex relationships from high-dimensional data.
  • Independent Component Analysis (ICA) is a statistical technique for separating independent sources from mixed signals.

Purpose of the Study:

  • To introduce and validate a novel method, Independent Component Analysis-Multiple Linear Regression (ICA-MLR), for QSPR analysis.
  • To compare the performance of ICA-MLR against established methods like PLS using simulation and real-world data.
  • To assess the ability of ICA-MLR to identify significant molecular descriptors contributing to aqueous solubility.

Related Experiment Videos

Main Methods:

  • Independent Component Analysis (ICA) was employed to extract mutually independent components from molecular descriptors.
  • Multiple Linear Regression (MLR) was used to establish a relationship between the extracted independent components and the objective variable (aqueous solubility).
  • The performance of the developed ICA-MLR model was compared with PLS and Genetic Algorithm-PLS models using metrics like R², Q², and Rpred².

Main Results:

  • ICA-MLR demonstrated superior predictive accuracy compared to PLS, evidenced by higher R² (0.937 vs. 0.836), Q² (0.868 vs. 0.819), and Rpred² (0.894 vs. 0.848) values.
  • The ICA-MLR model successfully identified significant components and their contributions to aqueous solubility through regression coefficients.
  • Simulation data confirmed the superiority of ICA-MLR over PLS in component extraction and predictive modeling.

Conclusions:

  • ICA-MLR is a powerful and effective method for QSPR analysis, offering improved predictive performance over traditional techniques.
  • The method successfully extracts relevant components from molecular descriptors, leading to highly accurate regression models.
  • ICA-MLR provides insights into the contribution of individual molecular descriptors to the analyzed property, enhancing model interpretability.