Title: | Machine learning based code smell detection |
Subject: | Computer science, Software engineering |
Level: | Advanced |
Description: |
Code smells refers to the design and/or implementation of source code that often lead to frequent changes and lead to software faults. A number of code smell detectors are developed with varying accuracy. However, a number of limitations still exist such as : (i) subjective preference of developers with respect to code smells detected by the tools, (ii) lack of agreement between different detectors, and (iii) difficulties in finding good thresholds to be used for detection. In this thesis, an investigation into the existing techniques are to be performed as well as develop ML techniques to overcome some of the limitations. Supervisor(s): Abu Naser Masud Examiner: |
Start date: | |
End date: | |
Prerequisites: | Knowledge of machine learning, software engineering, and tensor flow/keras are required. |
IDT supervisors: | |
Examiner: | |
Comments: | |
Company contact: |