TODO-List
Below is a to-do list for the LAD-WEKA project. The purpose of this list is twofold: (i) to provide an overview of the upcoming features that shall be included in the next releases of LAD-WEKA, and (ii) to encourage users of LAD-WEKA -- and members of the machine learning scientific community in general -- to get involved and contribute high-quality code to the LAD-WEKA project.
The following list is non-exhaustive, but includes the main targets for short-term improvement of LAD-WEKA:
The following list is non-exhaustive, but includes the main targets for short-term improvement of LAD-WEKA:
- Write code for the Maximized Spanned Patterns algorithm (from the Maximum Patterns in Datasets paper).
- Write a more scalable algorithm for set covering (to be used in the Greedy Set Cover approach for cutpoint selection). For instance, we can use one of the algorithms of P. Galinier and A. Hertz in "Solution Techniques for the Large Set Covering Problem".
- Provide support to classification problems containing multiple classes.