A professional-grade R machine learning workspace designed to run publication-ready caret, tidymodels, random forest, and XGBoost pipelines.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install r-machine-learning
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install r-machine-learning using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is R Machine Learning Workbench?
The R Machine Learning Workbench is an end-to-end machine learning execution platform tailored for biomedical researchers, bioinformaticians, and data scientists. Built natively in R, this workbench standardizes complex machine learning pipelines. It guides users seamlessly from raw data preprocessing and high-dimensional feature selection to predictive model optimization and explainability analysis without requiring manually written boilerplate scripts.\n\nBy implementing this workbench as part of the library of Openclaw Skills, researchers can automate rigorous model-building processes. The workbench provides integrated workflows for caret and tidymodels, supporting baseline and complex algorithms like Random Forests, XGBoost, LASSO, and Support Vector Machines. It eliminates manual intervention by auto-generating vector-based publication charts, comprehensive performance evaluations, and structured HTML reports.
R Machine Learning Workbench Use Cases
- Bioinformatics & Clinical Research: Train machine learning models for differential gene expression classification, survival analysis, and clinical outcome predictions.\n- High-Dimensional Feature Selection: Reduce features in high-dimensional sparse datasets (p >> n) using robust methods like Boruta, LASSO regression, and Recursive Feature Elimination (RFE).\n- Rigorous Cross-Validation: Execute systematic 5-fold or 10-fold cross-validation to assess model generalizability and avoid overfitting on small datasets.\n- Model Comparison & Statistical Testing: Compare multiple algorithms simultaneously to select the optimal model based on statistical performance metrics.\n- Explainable Artificial Intelligence (XAI): Demystify complex machine learning models using SHAP (Shapley Additive exPlanations) to discover and analyze feature contributions.
How R Machine Learning Workbench Works
- System Initialization: Run the dedicated installation script to automatically download, install, and configure required CRAN packages including caret, tidymodels, Boruta, and DALEX.\n2. Data Partitioning: Split input CSV datasets into clean training, validation, and testing subsets based on user-defined split ratios.\n3. Feature Selection: Run algorithms like Boruta, RFE, or LASSO to identify the most predictive feature subsets and eliminate noise.\n4. Model Optimization: Train classifiers or regression models while executing hyperparameter optimization using grid search, random search, or Bayesian optimization workflows.\n5. Validation & Assessment: Evaluate trained models using cross-validation alongside diagnostic techniques like ROC-AUC curves and calibration curves.\n6. Explainability Generation: Apply SHAP analysis to the final model parameters to calculate feature importances and directional impacts on predictions.\n7. Reporting: Compile final model configurations, prediction tables, vector plots, and detailed summaries into a standalone, interactive HTML report.
R Machine Learning Workbench Setup
Ensure that R (version 4.0 or higher) is installed on your system. Windows users must add the R executable path (e.g., C:\\Program Files\\R\\R-4.x.x\\bin\\x64) to their system Environment Variables (PATH).\n\n### Linux & macOS Installation and Execution\nbash\n# 1. Install all R package dependencies\nbash scripts/launcher.sh install\n\n# 2. Run the end-to-end machine learning pipeline\nbash scripts/launcher.sh pipeline --data data.csv --target outcome --type classification --cv 5\n\n# 3. Optimize models with Bayesian hyperparameter tuning\nbash scripts/launcher.sh tune --data data.csv --target outcome --model xgboost --method bayesian\n\n\n### Windows PowerShell Installation and Execution\npowershell\n# 1. Install all R package dependencies\n.\\scripts\\launcher.ps1 install\n\n# 2. Run the end-to-end pipeline\n.\\scripts\\launcher.ps1 pipeline --data data.csv --target outcome --type classification --cv 5\n\n# 3. Custom feature selection followed by Random Forest training\n.\\scripts\\launcher.ps1 feature_select --data train.csv --target outcome --method boruta\n.\\scripts\\launcher.ps1 rf --data train.csv --target outcome --cv 5 --tune TRUE\n
R Machine Learning Workbench Data Schema & Taxonomy
Input Dataset Specifications\nInput datasets must be provided as a standard CSV file where rows represent observation samples and columns represent features or variables. Ensure column names contain no special characters and missing values are represented as NA.\n\n| Pipeline Target Type | Target Variable Constraints | Required Fields / Format |\n| :--- | :--- | :--- |\n| Classification | Binary (0/1 or factor levels) or Multi-class | Structured factor variable target column |\n| Regression | Continuous numeric targets | Numerical values column |\n| Survival Analysis | Time-to-event outcomes | Distinct time (numeric) + status (0/1 binary event) columns |\n\n### Output Artifact Directory Structure\nAll results generated by the workbench are saved in an organized output/ directory format:\n\n| Directory | File Format | Contents |\n| :--- | :--- | :--- |\n| output/plots/ | .pdf, .png, .svg | Publication-ready plots (ROC curves, PR curves, SHAP plots, Calibration curves) |\n| output/models/ | .rds | Serialized R model objects ready for deployment, prediction, or reuse |\n| output/predictions/ | .csv | Model predictions on test datasets mapping raw records to class probabilities or values |\n| output/reports/ | .html, .pdf | Comprehensive, compiled evaluation reports generated via RMarkdown templates |\n| output/tables/ | .csv, .tex | Parameter logs, variable importance metrics, and model evaluation metrics for LaTeX integration |
R Machine Learning Workbench Advanced Features
- Survival Machine Learning Support: Built-in estimators for clinical survival data, incorporating Random Survival Forests (via randomForestSRC) and Cox proportional hazard XGBoost models.\n- Bayesian Hyperparameter Search: Automate hyperparameter optimization using Bayesian active search algorithms to identify optimal model weights faster than grid search.\n- Explainable AI Pipeline: Dynamic integration with the DALEX and fastshap ecosystems, automatically outputting explainable AI summaries and visual SHAP plots.\n- Multi-Model Evaluation & Selection: Directly train multiple model families simultaneously and run standardized statistical checks to programmatically select the champion model.
SKILL.md
Loading
Related Openclaw Skills
Claude Usage Companion
A hands-off reporting and reminder utility designed to monitor separate Claude programmatic credit pools and optimize working day interactive windows.
Grocery Price Comparer
An AI-powered grocery utility that normalizes units and compares fresh food prices across Hema, Dingdong, Meituan Maicai, and Pupu.
Shenzhen Zhongkao Enrollment Assistant
An AI-powered enrollment assistant that guides students and parents through the entire Shenzhen high school entrance exam (Zhongkao) process, from policy lookups to specialized talent admission strategies.
weekly-report
An AI-powered summary generator that transforms unstructured weekly logs into professional, quantified performance reports tailored to various corporate structures.
Legal Contract Reviewer
An AI-powered legal document audit tool designed to scan text, classify agreement types, flag high-risk terms, and supply actionable modification suggestions based on legal principles.
Website Cloner
Website Cloner enables 1:1 local replication of any online website while maintaining full structural, visual, and interactive fidelity.
Featured*
