smote defect prediction

Old data (procedural systems). Can mention this paper and say "in this paper we test if those limited results on older procedural systems hold for a wider range of new systems"

Repeat for all other papers that mention SMOTE:

respect them. mention them
disrespect them saying that their analysis was somehow limited and we need to do better. e.g. old data, not OO-data , limited data, complicated stats, antiquated states, etc etc .

want to see that list in your wed deliverable

btw: suggested title: "When making up data is a good idea: On the advantages of partially synthetic training sets for software analytics".

amritbhanu / fss16591

smote defect prediction #6