Michael Hanke
0bffdb7141
LVR and HHU are both in Düsseldorf, for example. This should be more stable regarding future extensions.
1.6 KiB
| title | contributors | sites | topics | weight | |||
|---|---|---|---|---|---|---|---|
| julearn: Machine learning for everyone |
|
|
|
1000 |
{{< lead >}} julearn is a user-oriented machine-learning library. It integrates machine learning workflows, including model assessment and comparison. The library enables users to effortlessly design and test machine learning models directly from pandas DataFrames, while maintaining the flexibility of utilizing scikit-learn's data processing tools and models.
{{< /lead >}}
Building, evaluating, reproducing and interpreting ML models from neuroimaging is not easy. julearn enables domain experts without highly developed programming and technical skills to analyze brain images and build complex ML pipelines, while neuroimaging and ML experts can easily extend the libraries with custom methods. At the same time, julearn prevents typical user errors, in particular bias caused by data leakage.
Advantages in brief:
- Minimal coding: Easily create and evaluate models
- Complex tasks made simple: Estimate model performance using cross-validation, easy hyperparameter tuning
- Robust: Made to prevent user-related errors like data-leakage
- Open and established: Built on top of state-of-the-art libraries (e.g. scikit-learn)
{{< icon "document" >}} Docs {{< icon "comment" >}} Support {{< icon "github" >}} GitHub
An overview of julearn is also available as an easily digestible leaflet: Download Flyer {{< icon "chevron-down" >}}