Predicting Open Source Software Sustainability with Deep Temporal Neural Hierarchical Architectures and Explainable AI

OSS sustainability is multidimensional, not just about project age: The research demonstrates that project health depends on multiple factors including contribution activity, community participation, and maintenance dynamics—rather than simply how long a project has existed. This suggests stakeholders should evaluate OSS projects using comprehensive health indicators rather than relying on single metrics.

Temporal patterns reveal project trajectories better than static snapshots: By analyzing 24-month activity sequences, the framework achieved over 94% accuracy in predicting lifecycle stages. This finding indicates that how contribution patterns change over time is more informative than aggregate statistics, suggesting organizations should monitor OSS projects continuously rather than conducting periodic reviews.

Explainable AI makes sustainability predictions transparent and actionable: The paper incorporates explainable AI techniques to show which factors drive predictions about project health. This transparency is crucial for responsible AI in critical infrastructure decisions—stakeholders can understand why a project is classified as sustainable or at-risk, enabling more informed decision-making about which projects to depend on or invest in.

Contribution activity is a key predictor of OSS sustainability: Attribution analyses consistently identified contribution patterns as central to the model's predictions. This empirical finding validates the intuition that active development communities are essential to project health, providing data-driven justification for organizations to prioritize supporting projects with engaged contributor bases.