What question did this study set out to answer?

Develop and evaluate machine learning models to accurately predict small-molecule compound solubility in various organic solvents at different temperatures.

December 1, 2021Journal of CheminformaticsOpen Access

Prediction of small-molecule compound solubility in organic solvents by machine learning algorithms

View Full Paper

Ask AI

Bookmark

Authors

ZYZhuyifan Ye DODefang Ouyang

Discussion

Member takes

Overview

Computational model evaluation demonstrates accurate small-molecule solubility prediction across organic solvents using LightGBM, suggesting an efficient framework for chemical solvent screening.

Key Points

Develop and evaluate machine learning models to accurately predict small-molecule compound solubility in various organic solvents at different temperatures.
Extracted and standardized an experimental dataset containing 5,081 temperature and solubility data points for organic solvent mixtures.
Used molecular fingerprints to capture structural features of small-molecule solutes.
Compared lightGBM against deep learning and traditional machine learning algorithms (PLS, Ridge regression, kNN, DT, ET, RF, and SVM).
LightGBM achieved superior overall generalization across temperatures compared to other models, yielding an error of logS ± 0.20.
For unseen solutes, LightGBM maintained a prediction accuracy of logS ± 0.59, matching the expected noise level of experimental solubility data.

Cite This Study

Ye et al. (2021) studied this question.

synapsesocial.com/papers/6a8871cb8dfe8b7ec5a548c1 https://doi.org/10.1186/s13321-021-00575-3

View Full Paper

Ask AI

Bookmark

December 1, 2021Journal of CheminformaticsOpen Access

Prediction of small-molecule compound solubility in organic solvents by machine learning algorithms

View Full Paper

Ask AI

Bookmark

Authors

ZYZhuyifan Ye DODefang Ouyang

Discussion

Member takes

Overview

Computational model evaluation demonstrates accurate small-molecule solubility prediction across organic solvents using LightGBM, suggesting an efficient framework for chemical solvent screening.

Key Points

Develop and evaluate machine learning models to accurately predict small-molecule compound solubility in various organic solvents at different temperatures.
Extracted and standardized an experimental dataset containing 5,081 temperature and solubility data points for organic solvent mixtures.
Used molecular fingerprints to capture structural features of small-molecule solutes.
Compared lightGBM against deep learning and traditional machine learning algorithms (PLS, Ridge regression, kNN, DT, ET, RF, and SVM).
LightGBM achieved superior overall generalization across temperatures compared to other models, yielding an error of logS ± 0.20.
For unseen solutes, LightGBM maintained a prediction accuracy of logS ± 0.59, matching the expected noise level of experimental solubility data.

Cite This Study

Ye et al. (2021) studied this question.

synapsesocial.com/papers/6a8871cb8dfe8b7ec5a548c1 https://doi.org/10.1186/s13321-021-00575-3

Also Consider

Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark