For LN: I have a rig to test multigoal learning. Combines Site and Tag prediction.
The idea here is to use tag-prediction results to correctly identify which site a post belongs to.
A new SVM is trained for each tag.
Doing
2 things:
For research: Distinguishing Cause from Effect in SE. Exploring 2 ideas:
The idea here is to see if changes introduced by XTREEs change/respect underlying cause-effect relationships in data. Ideal result: No! Current Result: Inconclusive, high variance.
Experimental: Exploring cause-effect relationships between static code metrics & anti-patterns. I'm creating a new dataset with PTIDEJ tool to do this.
To Do
Continue exploring XTREEs with cause effect. My results are inconclusive and unstable at the moment.
Results for LN as an issue report. Will have this before our meet tomorrow!
Meeting Manny today to help with some hackathon stuff.
Done
Doing
2 things:
To Do
Link to my cause-effect test -> link