One-click install
npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill aeon-silverstein
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: aeon
Source: https://github.com/silverstein/claude-scientific-skills-desktop/tree/main/corpus/aeon
Command: npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill aeon-silverstein

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you build accurate machine learning solutions for time-indexed data where patterns depend on order, timing, and temporal alignment.

Core Features & Use Cases

  • Time Series Modeling for Common Tasks: Use aeon to tackle classification, regression, clustering, forecasting, anomaly detection, segmentation, and similarity search on univariate or multivariate time series.
  • Scikit-learn Compatible Workflow: Supports familiar fit/predict and fit/transform patterns so you can integrate time series models into broader ML pipelines.
  • Practical Algorithm Selection Guidance: Provides recommendations for speed vs. accuracy, distance metrics (e.g., DTW), and feature extraction approaches (e.g., ROCKET/Catch22).

Use Case: You have sensor signals from multiple machines and need to detect anomalies and segment operating regimes; use aeon’s anomaly detection and segmentation methods to find unusual subsequences and change points.

Quick Start

Use the aeon skill to train a time-series classifier on your dataset and return predictions for the test split by identifying an appropriate estimator (e.g., a ROCKET variant) based on your speed/accuracy needs.

Frequently Asked Questions about aeon

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a time-series classification model for multivariate sensor data?

To build a time-series classification model, you can use scikit-learn compatible fit/predict patterns on your multivariate temporal data. Tools like ROCKET variants provide recommendations for balancing speed and accuracy during model training and prediction.

What is the best way to detect anomalies and segment operating regimes in time series data?

Anomaly detection and segmentation identify unusual subsequences and change points in your temporal data. You can apply these methods to sequential sensor signals to find distinct operating regimes and flag unusual patterns.

Can I use aeon with my existing scikit-learn ML pipelines?

Yes, aeon satisfies scikit-learn compatible APIs, supporting familiar fit/predict and fit/transform patterns. This allows you to seamlessly integrate time-series models into your broader machine learning workflows.

How does matrix-profile-style motif and discord discovery work for time series?

Matrix-profile-style discovery finds repeating patterns (motifs) and anomalies (discords) within sequential temporal data. It leverages distance metrics and similarity search to identify these meaningful subsequences in univariate or multivariate series.

When should I choose DTW distance metrics versus feature extraction like Catch22 for time series forecasting?

Choose DTW distance metrics for temporal shift alignment or Catch22 feature extraction for fast accuracy. Algorithm selection guidance helps you trade off speed versus accuracy based on your specific forecasting or clustering needs.