A newly developed hybrid software fault prediction framework achieves superior predictive capability and stronger generalization than models relying solely on conventional metrics by incorporating design-level violation measurements. Designed to improve software reliability and decrease maintenance expenditures, the framework addresses long-standing limitations in defect estimation.
Standard fault prediction approaches depend primarily on traditional metrics, encompassing object-oriented, process, and complexity measures. While these conventional metrics successfully capture structural and historical properties across source code repositories, they frequently fail to detect design-level violations that ultimately contribute to software faults. To bridge this critical gap, the investigation introduces a hybrid framework that integrates established metrics with newly formulated SOLID principle-based design metrics.
At the foundation of this hybrid framework are two specific design violation metrics: the Liskov Substitution Principle Violation metric, designated LSPV, and the Unified Interface Segregation Principle metric, designated UISP. These two metrics mathematically quantify behavioral inconsistencies, interface misuse, and unnecessary coupling across object-oriented architectures.
To process these features efficiently, the framework applies Principal Component Analysis to reduce redundancy and eliminate multicollinearity among the metric inputs. The framework then combines the LSPV and UISP components into an integrated measure known as the Hybrid Design Violation Metric, or HDVM. This metric employs a sigmoid-based mathematical formulation to probabilistically combine the design violation indicators and estimate component-level fault-proneness.
Because existing software engineering benchmark datasets do not incorporate SOLID principle-based design metrics, the research required the construction of a dedicated empirical repository. The authors assembled a new dataset gathered from three major open-source Java projects: Apache Ant, Apache Lucene, and Apache Camel. To conduct a rigorous evaluation of HDVM, the team additionally developed a controlled mutation-based dataset. This synthetic testbed was constructed by systematically injecting representative Liskov Substitution Principle and Interface Segregation Principle violations into the codebase while strictly preserving system functionality and compilability.
Comprehensive experimental results and ablation analysis revealed that models utilizing the integrated HDVM metric consistently achieve superior predictive performance compared to models dependent entirely on conventional metrics.
Furthermore, specialized bug-type analysis demonstrated that HDVM is exceptionally effective in pinpointing software faults stemming from inheritance misuse, polymorphism violations, interface inconsistencies, and architectural coupling degradation.
Subject of Research: Software fault prediction using SOLID principle design violation metrics
Article Title: Design-level software fault prediction using SOLID-based LSP and ISP violation metrics
Article References: Ray, M. (2026). Design-level software fault prediction using SOLID-based LSP and ISP violation metrics. Industrial Artificial Intelligence. https://doi.org/10.1007/s44244-026-00037-5
Image Credits: AI Generated
DOI: 10.1007/s44244-026-00037-5
Keywords: software fault prediction, SOLID principles, Liskov Substitution Principle, Interface Segregation Principle, LSPV metric, UISP metric, Hybrid Design Violation Metric, object-oriented systems, Principal Component Analysis, Apache Ant, Apache Lucene, Apache Camel, mutation-based testing, software reliability, polymorphism violations, inheritance misuse, architectural coupling, ablation analysis
Cite Scienmag News
Reid Dalton. (November 16, 2026). SOLID Design Violation Metrics Improve Fault Prediction in Java Systems. Scienmag. https://scienmag.com/solid-design-violation-metrics-improve-fault-prediction-in-java-systems/
Reid Dalton. "SOLID Design Violation Metrics Improve Fault Prediction in Java Systems." Scienmag, 16 November 2026, https://scienmag.com/solid-design-violation-metrics-improve-fault-prediction-in-java-systems/. Accessed 16 November 2026.
Reid Dalton. "SOLID Design Violation Metrics Improve Fault Prediction in Java Systems." Scienmag. November 16, 2026. https://scienmag.com/solid-design-violation-metrics-improve-fault-prediction-in-java-systems/








