analytics-ml

name

analytics-ml

description

Machine learning modeling conventions for applied analytics: feature engineering, model evaluation, validation strategy, and deployment considerations. Use this skill when building predictive models, selecting features, choosing evaluation metrics, setting up train/test splits, or discussing ML methodology for business applications. Also apply when the user asks about classification, regression, churn prediction, forecasting, or any supervised/unsupervised learning task in an analytics context. Use when this capability is needed.

ML Modeling

This skill is planned but not yet implemented.

Conventions for applied machine learning in analytics contexts: predicting outcomes, scoring users, forecasting metrics. This is the prescriptive/predictive domain, distinct from experimentation (which establishes causation through randomization).

When complete, this skill will cover:

• Feature engineering patterns for tabular business data
• Model evaluation: choosing the right metric for the business question (precision vs. recall, RMSE vs. MAE, AUC vs. F1)
• Validation strategy: time-based splits for temporal data, stratification, cross-validation
• Overfitting discipline: regularization, feature selection, holdout discipline
• Forecasting conventions: time series decomposition, seasonality, trend
• Deployment considerations: batch vs. real-time, monitoring, drift detection

References

• Anthropic Agent Skills: check for updated patterns and templates
• scikit-learn: the default ML library for tabular analytics work
• statsmodels: statistical modeling in Python

---

Source: aleberriz/agent-skills — distributed by TomeVault.