khiva ☆
« Back to VersTracker
Description:
Algorithms to analyse time series
Type: Formula  |  Tracked Since: Dec 28, 2025
Links: Homepage  |  formulae.brew.sh
Category: Ai ml
Tags: time-series data-science gpu machine-learning algorithms
Install: brew install khiva
About:
Khiva is an accelerated library providing efficient time-series analysis algorithms. It enables data scientists to detect patterns, shape features, and reduce dimensionality directly on the GPU. This allows for faster processing of large datasets compared to CPU-based libraries.
Key Features:
  • GPU-accelerated time-series analysis
  • Includes clustering, classification, and dimensionality reduction
  • Cross-language bindings (Python, C++, Java)
  • OpenCL backend for hardware portability
Use Cases:
  • Anomaly detection in IoT sensor data
  • Classification of industrial or financial time-series patterns
  • Feature extraction for machine learning pipelines
Alternatives:
  • tsfresh – Python library focused on feature extraction, runs on CPU.
  • sktime – Unified framework for time-series algorithms, CPU-based.
Version History
Detected Version Rev Change Commit
Sep 13, 2025 11:05am 0 VERSION_BUMP 3946cda7
Sep 27, 2024 11:01am 0 VERSION_BUMP 35adf72f
« Back to VersTracker  |  Browse Packages  |  About