An expert scikit-learn compatible Python toolkit for advanced time series machine learning tasks like forecasting, classification, clustering, and anomaly detection.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install aeon
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 aeon 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 aeon?
The aeon skill equips an AI coding agent with a specialized, scikit-learn compatible Python framework designed exclusively for temporal data, sequential patterns, and time-indexed observations. Unlike general-purpose machine learning libraries that flatten sequences and lose vital temporal contexts, aeon implements dedicated, state-of-the-art algorithms tailored specifically for complex univariate and multivariate time series analysis.
By adding this toolkit to an AI agent's repository, engineers gain an immediate pathway to sophisticated temporal modeling. It bridges traditional ML ecosystems and advanced deep learning architectures, providing seamless integration with existing scikit-learn workflows while solving deep technical problems in time-domain feature engineering.
aeon Use Cases
- Classifying or predicting categorical outcomes from structured time series sequences.
- Detecting anomalies, outliers, or structural change points in real-time or historical temporal data.
- Clustering similar unlabelled time series patterns to identify hidden behaviors.
- Forecasting future values across sequential data arrays using traditional and deep learning architectures.
- Executing similarity searches to isolate repeating motifs or highly unusual subsequences (discords).
- Computing specialized temporal distance matrices like Dynamic Time Warping (DTW) across variable-length series.
How aeon Works
- Data Ingestion and Validation: The skill formats input data into a 3D tensor structure of shape
(n_samples, n_channels, n_timepoints)required by specialized temporal estimators. - Preprocessing and Normalization: The pipeline executes z-normalization or mean imputation to ensure statistical scaling across unequal sequences.
- Feature Transformation: Applied feature extractors like ROCKET variants or Catch22 compile temporal dynamics into high-performance tabular snapshots.
- Model Application: The target algorithm performs forecasting, classification, regression, or clustering using scikit-learn compatible fit/predict APIs.
- Evaluation and Benchmarking: The output results are validated against built-in datasets or published benchmarks to gauge accurate baseline improvements.
aeon Setup
Install the core library securely inside your environment using the following command:
uv pip install aeon
aeon Data Schema & Taxonomy
Input Tensor Shape
aeon algorithms enforce strict data shapes to preserve multi-channel sequence intelligence:
| Dimension | Description | Context |
|---|---|---|
n_samples |
Count of unique series arrays | Rows in data matrix |
n_channels |
Feature streams or sensors per sample | Univariate (1) or Multivariate (>1) |
n_timepoints |
Explicit length of individual observation intervals | Steps along the sequence axis |
Internal Reference Map
The skill establishes structural documentation dependencies under the references/ directory to coordinate specialized algorithm behavior:
classification.md- Complete classification algorithm catalog.regression.md- Continuous estimation and regression definitions.clustering.md- Unsupervised grouping and centroid configurations.forecasting.md- Horizon predicting structures.anomaly_detection.md- Outlier scoring parameters.segmentation.md- Change point partitioning layouts.similarity_search.md- Pattern matching and motif configurations.transformations.md- Feature extraction and preprocessing mechanics.distances.md- Time series distance metrics (DTW, MSM, ERP).networks.md- Deep learning architectural configurations.datasets_benchmarking.md- Standard data loading formats.
aeon Advanced Features
- Advanced Temporal Distance Metrics: Native support for complex elastic, lock-step, and shape-based sequence distances including DTW, DDTW, ERP, and Shape DTW.
- Convolution-Based ROCKET Pipelines: High-speed, high-performance feature extraction utilizing random convolutional kernels (
MiniRocketClassifier,Arsenal) for efficient scaling. - Built-in Deep Learning Architectures: Direct access to deep neural networks engineered for sequences, such as
InceptionTimeClassifier,ResNetClassifier, andTCNNetworkmodels. - Scikit-Learn Native Pipeline Compatibility: Seamlessly combines specialized transformations with standard
sklearn.pipeline.Pipelineworkflows for unified model execution. - Benchmarking Engine: Built-in access to benchmark datasets with automated comparison functions against published academic baselines.
SKILL.md
Loading
Related Openclaw Skills
Doc2md
A high-performance, cross-platform CLI tool designed to batch-convert PDF, DOCX, and PPTX documents into Markdown using the doc2md API.
Adaptyv Bio Foundry
A dedicated integration designed to automate protein experiment design, submission, and assay results retrieval using the Adaptyv Bio Foundry API and Python SDK.
Antom Reconciliation Expert
An AI-powered reconciliation assistant that parses, validates, and analyzes local Settlement Detail report files while offering natural-language domain Q&A.
Undaunted Courage: Expedition History and Leadership Toolkit
An AI-powered historical toolkit that translates the history of the Lewis and Clark Expedition and Meriwether Lewis's leadership into actionable lessons.
Gut Health Assistant (Giulia Enders's Gut)
An AI-powered digestive health companion providing scientific insights into the microbiome, gut-brain axis, and daily wellness based on Giulia Enders's bestselling book.
Remotion Video Pipeline
An audio-driven video orchestration skill that guides users through a repeatable pipeline to produce programmatic Remotion short-form videos with human verification gates.
Featured*
