Open Suvodeep90 opened 4 years ago
What is the difference between GENERAL and hierarchical clustering other than the GENERAL is applied to defect prediction?
GENERAL is an algorithm/Framework to approximately find a bellwether project (source project) for transfer learning. GENERAL algorithm uses the hierarchical clustering to cluster the data into smaller clusters at the leaf level and then into larger clusters as you go up. Then we use a n*n comparison in each leaf cluster and then use an approximation method to identify bellwether at higher nodes.
What I find irritating is that the manuscript introduces a new approach GENERAL, which just applies hierarchical clustering. As far as I understand, hierarchical clustering is a well-known, standard, and largely-evaluated clustering algorithm. I don’t really see the need to rebrand it with a new name. What is the difference between GENERAL and hierarchical clustering other than the GENERAL is applied to defect prediction?