SEDiL
Software for Edit Distance Learning
SEDiL is a software platform created for the ANR Marmota Project. This platform aim to publicly give access to algorithm around the domain of edit distance between structured data (mainly sequences and tree).
The initiative of the creation of this platform is mainly due to Marc Sebban . The creation of this platform was the main goal of my post-PhD in the Hubert Curien Laboratory.
Edit distances are used in many domains such as bio-informatic, image, sound, music recognition, WEB...
However, in real usages, edit distances parameters are fixed manually. Most of time nobody has the necessary expertise to tune these values. This is why it seems usefull to learn the edit distances parameters.
Because XML is becoming the new standard for information storage. Providing more informations on informations than just raw text storage. Furthermore, trees are a generalisation of sequences and therefore, computing and learning tree edit distances gives immediately the capability to compute and learn sequence edit distances.
This is a way to gives the distance between two structured data giving the minimal number of necessary modifications (edit operations) to transform one to another.