Skip to content

SMILE is a feature‑based machine learning classification framework for event‑interval sequence data. While its primary application is medical diagnostics using Electronic Health Record data, the framework is generalisable to domains such as human activity recognition, music classification, and manufacturing telemetry.

License

Notifications You must be signed in to change notification settings

JonRebane/smile-ml-framework

Repository files navigation

SMILE

Purpose:

Implementation of the SMILE framework in the Research Project concerning Sequences of temporal Intervals. https://link.springer.com/article/10.1007/s10618-020-00719-3

Contribution:

Improves predictive performance (AUC) compared to state‑of‑the‑art distance‑based and feature‑based classifiers by incorporating both temporal relations and event durations.

Quickstart:

It was was developed as an eclipse project, so the simplest way to get it running is to import it in eclipse: File -> Import -> Git -> Projects from Git -> clone URI The experiments can then be executed by running the main methods in experiment.RuntimeComparisonRealDatasets and experiment.RuntimeComparisonSyntheticDatasets respectively. In order to successfully run the classification algorithms for the real datasets the java process should be given 4GB Main Memory since large data structures will bee needed (at least for the multi-labeled datasets ASL-BU and ASL-BU-2).

About

SMILE is a feature‑based machine learning classification framework for event‑interval sequence data. While its primary application is medical diagnostics using Electronic Health Record data, the framework is generalisable to domains such as human activity recognition, music classification, and manufacturing telemetry.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •