Detect Unknown Licenses and Indirect License references in ScanCode

Scancode uses various different techniques to detect licenses based on various multiple sequence alignments ,algorithms ,automations etc. but the detection results are not always accurate enough as expected. The aim of this project is to improve license detection accuracy in general. For example: Detecting unknown licenses and study indirect references and as a whole gain more insight into the Scan Results. Goals to be accomplished:

1)Main motive will be to improve license detection accuracy for the licenses that are declared as “Unknown” and optimize somehow the way they are handled currently.

2)Detailed review of the licenses which are marked as “Unknown License Reference”.

3)General Improvements for the licenses that are detected incorrectly .Having statistical analysis of all license detection scans (including minimum coverage , score etc.) and of existing license texts and rules.