RRegrs

RRegrs builds and evaluates predictive regression models within R by integrating multiple algorithms and standardized cross-validation workflows.

Key Features:

• Multiple Regression Algorithms: Implements Multiple Linear Regression, Generalized Linear Models with stepwise feature selection, Partial Least Squares Regression, Lasso Regression, and Support Vector Machines with Recursive Feature Elimination through the caret framework.
• Automated Cross-Validation and Reporting: Performs repeated 10-fold and leave-one-out cross-validation, automates parameter selection and performance evaluation, and generates standardized comparative reports.

Scientific Applications:

• QSAR and Cheminformatics Modeling: Supports quantitative structure–activity relationship analyses, proteomics datasets, nano-metal oxide descriptor modeling, and acute aquatic toxicity prediction.

Methodology:

RRegrs integrates caret-based regression workflows, applies multiple modeling algorithms with systematic parameter tuning, evaluates models using repeated 10-fold and leave-one-out cross-validation, and compares performance across standardized training and test datasets.