{"categories":[{"label":"Artificial Intelligence","url":"https://skillfed.io/packages/category/scientific-engineering-artificial-intelligence/9"},{"label":"Application Frameworks","url":"https://skillfed.io/packages/category/software-development-libraries-application-frameworks/5"}],"enrichment":{"capability":"BenchPOTS provides standardized preprocessing pipelines and evaluation tasks for benchmarking machine learning algorithms on partially-observed time series datasets.","skillfed_tags":["time-series-ml","benchmarking","missing-data"],"use_cases":["Evaluate imputation algorithms on standard POTS datasets with controlled missing-data rates.","Benchmark time-series forecasting models on partially-observed data from medical or sensor domains.","Compare classification and clustering performance across multiple incomplete time-series datasets.","Preprocess and standardize POTS datasets for reproducible machine learning research.","Validate new time-series algorithms against established benchmarks before publication."],"what_it_does":"BenchPOTS is a benchmarking toolkit for evaluating machine learning algorithms on partially-observed time series (POTS)\u2014datasets with missing or irregularly sampled values. It provides unified preprocessing pipelines for standard POTS datasets and a suite of evaluation tasks to measure algorithm performance on problems like imputation, classification, clustering, and forecasting.\n\nThe package wraps datasets like PhysioNet2012 and applies configurable missing-data patterns (e.g., MCAR with specified rates) to create reproducible benchmarks. It depends on torch, scikit-learn, pandas, numpy, and h5py for core functionality, plus specialized time-series libraries (tsdb, pygrinder, nonlinear_benchmarks). It targets researchers and practitioners developing or comparing algorithms for incomplete time series.","worth_installing":"Yes, if you are developing or comparing machine learning algorithms for partially-observed time series. The package offers low install friction, active maintenance, permissive licensing, and no known vulnerabilities. Install with caution if you have limited disk or memory\u2014the 8 runtime dependencies (especially torch) are substantial. Not necessary for one-off time-series tasks; best suited for research and systematic benchmarking."},"id":"benchpots","links":{"html":"https://skillfed.io/packages/benchpots","md":"https://skillfed.io/packages/benchpots.md","pypi":"https://pypi.org/project/benchpots/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2026-04-27","license_spdx":null,"license_treatment":"permissive","name":"benchpots","python_support":"supports_current","summary":"A Python Toolbox for Benchmarking Machine Learning on Partially-Observed Time Series"},"popularity":{"monthly_downloads":122414,"position":11952,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"1"}