Related Experiment Videos
PVLOO-Based Training Set Selection Improves the External Predictability of QSAR/QSPR Models.
Ying Dong1, Bingren Xiang2, Ding Du1
1Department of Organic Chemistry, College of Science, China Pharmaceutical University , 639 Longmian Avenue, Nanjing 211198, P. R. China.
Journal of Chemical Information and Modeling
|April 19, 2017
Summary
Selecting training sets using posterior variance of leave-one-out cross-validation (PVLOO) from Gaussian process (GP) models enhances external predictability in QSAR/QSPR studies. This method improves model performance over traditional division techniques.
Area of Science:
- Quantitative Structure-Activity Relationships (QSAR)
- Quantitative Structure-Property Relationships (QSPR)
- Cheminformatics
- Machine Learning
Background:
- Statistical external validation is crucial for assessing QSAR/QSPR model predictability.
- Training set selection significantly impacts model performance.
- Gaussian processes (GP) offer a probabilistic framework for modeling.
Purpose of the Study:
- Develop a novel training set division method for QSAR/QSPR.
- Improve the external predictability of predictive models.
- Investigate the utility of posterior variance of leave-one-out cross-validation (PVLOO) in Gaussian processes for training set selection.
Main Methods:
- Developed a division algorithm based on GP's PVLOO.
- Collected and analyzed four diverse, high-quality chemical data sets.
- Compared PVLOO-based selection with Kennard-Stone and random division.
- Utilized squared exponential, rational quadratic, and neural network covariance functions.
Main Results:
- Training sets selected using higher PVLOO values demonstrated statistically superior external predictability.
- PVLOO-based selection outperformed Kennard-Stone and random division methods.
- Root Mean Squared Error (RMSE) of external validation was the primary comparison metric.
Conclusions:
- PVLOO of Gaussian processes can effectively guide training set selection for enhanced QSAR/QSPR model predictability.
- Higher PVLOO values may correlate with greater mechanism diversity within training compounds.
- This approach offers a statistically robust method for optimizing predictive model development.