-
Notifications
You must be signed in to change notification settings - Fork 11
Features weights learning
André Freitas edited this page May 22, 2015
·
4 revisions
In Schwa we use fixes, authors and revisions metrics to estimate defect probabilities in Software components(files, classes, methods). The problem is that we need to find the weights of these features.
Using Genetic Algorithms, we found a way of computing these weights. The problem is that we need more computational power and diversity in repositories. A good way is doing "crowdsourcing". Please do the following:
- Install Schwa by reading the install instructions
- Download this script and run by giving a path of a Git repository you have:
chmod +x ./schwa_learning.sh
./schwa_learning.sh git/repository
We do not collect any sensitive information.
This will take a while and we will be thankful for your help :)