MEX: Metadata for Machine Learning Outputs
This project aims at facilitating the sharing and data management of machine learning experiment outputs.
For more information, please see the sub projects listed bellow:
- MEX Vocabulary: A Lightweight Interchange Format for Machine Learning Experiments
- LOG4MEX Java Library
- MEX Interfaces
- WASOTA Metadata Repository
- Get in touch! Google Group
##LOG4MEX Library: Demo and Documentation
##License MEX is Open Source and licensed under the GNU General Public License 3. (Copyright (c) 2014-2016, Diego Esteves)
MEX uses several other libraries. An incomplete list is as follows:
- Google Guava (Apache License Version 2.0)
- Log4j (Apache License Version 2.0)
- Apache Jena (Apache License Version 2.0)
For semantic web users the usage is straightforward, once the vocabulary can easily be handled with semantic web technologies, such as Apache Jena.
For non-semantic web users, there are two ways to use MEX
: if you're coding, you can import LOG4MEX, which implements an interface to be consumed into the user code for directly exporting the metadata in a simple manner, regardless possible semantic web aspects.
More recently, we have started the Framework's development, which aims at automatizing the task of generating metadata for Java
classes based on the concepts of reflection
, annotation
and generics
.
Finally, soon you'll be able to benefit from a REST interface and export your metadada via HTTP.