Evaluating High-Variance Leaves as Uncertainty Measure for Random Forest Regression

Machine learning for skin permeability prediction: random forest and XG boost regression

Background: Machine learning algorithms that can quickly and easily estimate skin permeability (Kp) are increasingly being used in drug delivery research. The linear free energy relationship (LFER) developed by Abraham is a practical technique for predicting Kp. The permeability coefficients and Abraham solute descriptor values for 175 organic compounds have been documented in the scientific literature.Purpose: The purpose of this project was to use a publicly available dataset to make skin permeability predictions using the random forest and XBoost regression techniques.Methods: We employed Pandas-based methods in JupyterLab to predict permeability coefficient (Kp) from solute descriptors (excess molar refraction [E], combined dipolarity/polarizability [S], overall solute hydrogen bond acidity and basicity [A and B], and the McGowan's characteristic molecular volume [V]).Results: The random forest and XG Boost regression models established statistically significant association between the descriptors and the skin permeability coefficient.

Chemoinformatic regression methods and their applicability domain

The growing interest in chemoinformatic model uncertainty calls for a summary of the most widely used regression techniques and how to estimate their reliability. Regression models learn a mapping from the space of explanatory variables to the space of continuous output values. Among other limitations, the predictive performance of the model is restricted by the training data used for model fitting. Identification of unusual objects by outlier detection methods can improve model performance. Additionally, proper model evaluation necessitates defining the limitations of the model, often called the applicability domain. Comparable to certain classifiers, some regression techniques come with built-in methods or augmentations to quantify their (un)certainty, while others rely on generic procedures. The theoretical background of their working principles and how to deduce specific and general definitions for their domain of applicability shall be explained.

Molecular Hessian matrices from a machine learning random forest regression algorithm

In this article, we present a machine learning model to obtain fast and accurate estimates of the molecular Hessian matrix. In this model, based on a random forest, the second derivatives of the energy with respect to redundant internal coordinates are learned individually. The internal coordinates together with their specific representation guarantee rotational and translational invariance. The model is trained on a subset of the QM7 dataset but is shown to be applicable to larger molecules picked from the QM9 dataset. From the predicted Hessian, it is also possible to obtain reasonable estimates of the vibrational frequencies, normal modes, and zero point energies of the molecules.

Large-scale evaluation of k-fold cross-validation ensembles for uncertainty estimation

It is insightful to report an estimator that describes how certain a model is in a prediction, additionally to the prediction alone. For regression tasks, most approaches implement a variation of the ensemble method, apart from few exceptions. Instead of a single estimator, a group of estimators yields several predictions for an input. The uncertainty can then be quantified by measuring the disagreement between the predictions, for example by the standard deviation. In theory, ensembles should not only provide uncertainties, they also boost the predictive performance by reducing errors arising from variance. Despite the development of novel methods, they are still considered the "golden-standard" to quantify the uncertainty of regression models. Subsampling-based methods to obtain ensembles can be applied to all models, regardless whether they are related to deep learning or traditional machine learning. However, little attention has been given to the question whether the ensemble method is applicable to virtually all scenarios occurring in the field of cheminformatics. In a widespread and diversified attempt, ensembles are evaluated for 32 datasets of different sizes and modeling difficulty, ranging from physicochemical properties to biological activities. For increasing ensemble sizes with up to 200 members, the predictive performance as well as the applicability as uncertainty estimator are shown for all combinations of five modeling techniques and four molecular featurizations. Useful recommendations were derived for practitioners regarding the success and minimum size of ensembles, depending on whether predictive performance or uncertainty quantification is of more importance for the task at hand.

Correlation and predictive ability of sensory characteristics and social interaction in children with autism spectrum disorder

Background

Individuals with autism spectrum disorder (ASD) often have different social characteristics and particular sensory processing patterns, and these sensory behaviors may affect their social functioning. The objective of our study is to investigate the sensory profiles of children with ASD and their association with social behavior. Specifically, we aim to identify the predictive role of sensory processing in social functioning.

Methods

The Short Sensory Profile (SSP) was utilized to analyze sensory differences between ASD children and their peers. The Social Responsiveness Scale (SRS) and other clinical scales were employed to assess the social functioning of children with ASD. Additionally, the predictive ability of sensory perception on social performance was discussed using random forest and support vector machine (SVM) models.

Results

The SSP scores of ASD children were lower than those of the control group, and there was a significant negative correlation between SSP scores and clinical scale scores (P < 0.05). The random forest and SVM models, using all the features, showed higher sensitivity, while the random forest model with 7-feature factors had the highest specificity. The area under the receiver operating characteristic (ROC) curve (AUC) for all the models was higher than 0.8.

Conclusion

Autistic children in our study have different patterns of sensory processing than their peers, which are significantly related to their patterns of social functioning. Sensory features can serve as a good predictor of social functioning in individuals with ASD.

Analysis of ecosystem service drivers based on interpretive machine learning: a case study of Zhejiang Province, China

A systematic understanding of the driving mechanisms of ecosystem services (ESs) and the relationships among them is critical for successful ecosystem management. However, the impact of driving factors on the relationships between ESs and the formation of ecosystem service bundles (ESBs) remains unclear. To address this gap, we developed a modeling process that used random forest (RF) to model the ESs and ESBs of Zhejiang Province, China, in regression and classification mode, respectively, and the Shapley Additive Explanations (SHAP) method to interpret the underlying driving forces. We first mapped the spatial distribution of seven ESs in Zhejiang Province at a 1 × 1 km spatial resolution and then used the K-means clustering algorithm to obtain four ESBs. Combining the RF models with SHAP analysis, the results showed that each ES had key driving factors, and the relationships of synergy and trade-off between ESs were determined by the driving direction and intensity of the key factors. The driving factors affect the relationships of ESs and consequently affect the formation of ESBs. Thus, managing the dominant drivers is key to improving the supply capacity of ESs.