Abstract
Evolvability defects are non-understandable and non-modifiable states that do not directly produce runtime behavioral failures. Automatic source code evaluation by metrics and thresholds can help reduce the burden of a manual inspection. This study addresses two problems. (1) Evolvability defects are not usually managed in bug tracking systems. (2) Conventional methods cannot fully interpret the relations among the metrics in a given context (e.g., programming language, application domain). The key actions of our method are to (1) gather trainingdata for machine learning by experts' manual inspection of some of the files in given systems (benchmark) and (2) employ a classification-tree learner algorithm, C5.0, which can deal with non-orthogonal relations between metrics. Furthermore, we experimentally confirm that, even with less training-data, our method provides a more precise evaluation than four conventional methods (the percentile, Alves' method, Bender's method, and the ROC curve-based method).
| Original language | English |
|---|---|
| Title of host publication | Proceedings - 2018 IEEE 18th International Conference on Software Quality, Reliability, and Security, QRS 2018 |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 83-94 |
| Number of pages | 12 |
| ISBN (Print) | 9781538677575 |
| DOIs | |
| Publication status | Published - 2018 Aug 2 |
| Event | 18th IEEE International Conference on Software Quality, Reliability, and Security, QRS 2018 - Lisbon, Portugal Duration: 2018 Jul 16 → 2018 Jul 20 |
Other
| Other | 18th IEEE International Conference on Software Quality, Reliability, and Security, QRS 2018 |
|---|---|
| Country | Portugal |
| City | Lisbon |
| Period | 18/7/16 → 18/7/20 |
Fingerprint
Keywords
- Classification-tree
- Contextual thresholds
- Evolvability defects
- Goal-Question-Metrics (GQM)
- Machine learning
- Software metrics
ASJC Scopus subject areas
- Software
- Safety, Risk, Reliability and Quality
Cite this
Machine learning to evaluate evolvability defects : Code metrics thresholds for a given context. / Tsuda, Naohiko; Washizaki, Hironori; Fukazawa, Yoshiaki; Yasuda, Yuichiro; Sugimura, Shunsuke.
Proceedings - 2018 IEEE 18th International Conference on Software Quality, Reliability, and Security, QRS 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 83-94 8424960.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Machine learning to evaluate evolvability defects
T2 - Code metrics thresholds for a given context
AU - Tsuda, Naohiko
AU - Washizaki, Hironori
AU - Fukazawa, Yoshiaki
AU - Yasuda, Yuichiro
AU - Sugimura, Shunsuke
PY - 2018/8/2
Y1 - 2018/8/2
N2 - Evolvability defects are non-understandable and non-modifiable states that do not directly produce runtime behavioral failures. Automatic source code evaluation by metrics and thresholds can help reduce the burden of a manual inspection. This study addresses two problems. (1) Evolvability defects are not usually managed in bug tracking systems. (2) Conventional methods cannot fully interpret the relations among the metrics in a given context (e.g., programming language, application domain). The key actions of our method are to (1) gather trainingdata for machine learning by experts' manual inspection of some of the files in given systems (benchmark) and (2) employ a classification-tree learner algorithm, C5.0, which can deal with non-orthogonal relations between metrics. Furthermore, we experimentally confirm that, even with less training-data, our method provides a more precise evaluation than four conventional methods (the percentile, Alves' method, Bender's method, and the ROC curve-based method).
AB - Evolvability defects are non-understandable and non-modifiable states that do not directly produce runtime behavioral failures. Automatic source code evaluation by metrics and thresholds can help reduce the burden of a manual inspection. This study addresses two problems. (1) Evolvability defects are not usually managed in bug tracking systems. (2) Conventional methods cannot fully interpret the relations among the metrics in a given context (e.g., programming language, application domain). The key actions of our method are to (1) gather trainingdata for machine learning by experts' manual inspection of some of the files in given systems (benchmark) and (2) employ a classification-tree learner algorithm, C5.0, which can deal with non-orthogonal relations between metrics. Furthermore, we experimentally confirm that, even with less training-data, our method provides a more precise evaluation than four conventional methods (the percentile, Alves' method, Bender's method, and the ROC curve-based method).
KW - Classification-tree
KW - Contextual thresholds
KW - Evolvability defects
KW - Goal-Question-Metrics (GQM)
KW - Machine learning
KW - Software metrics
UR - http://www.scopus.com/inward/record.url?scp=85052334218&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85052334218&partnerID=8YFLogxK
U2 - 10.1109/QRS.2018.00022
DO - 10.1109/QRS.2018.00022
M3 - Conference contribution
AN - SCOPUS:85052334218
SN - 9781538677575
SP - 83
EP - 94
BT - Proceedings - 2018 IEEE 18th International Conference on Software Quality, Reliability, and Security, QRS 2018
PB - Institute of Electrical and Electronics Engineers Inc.
ER -