Many organizations depend on software systems to accomplish their daily tasks, but these systems need to be maintained and evolved to cope with various changes and requirements. Before starting to maintain and evolve software systems, it is necessary to understand them. Reverse engineering plays a crucial role in comprehending various aspects of software systems by extracting different models and diagrams that represent the structure and behaviour of software systems. This article presents a systematic literature review (SLR) to understand the current state of research in model-driven engineering (MDE) for reverse engineering software systems. The considered articles came from five electronic databases (Scopus, IEEE Xplore, Web of Science, ACM Digital Library, and Google Scholar), and were supplemented by additional articles recommended by experts and provided by manual snowballing. From 538 surveyed papers, 83 principal studies were selected, which present the main characteristics of 64 model-driven reverse engineering (MDRE) approaches. These approaches are analyzed and evaluated based on their objectives and characteristics. Additionally, research gaps and areas where more research is needed are also identified. Therefore, the review provides comprehensive answers to several widely interesting questions for researchers and practitioners who are considering using MDRE.

Effective student management is crucial for fostering productive learning environments. This study presents a hybrid framework integrating machine learning (ML) techniques with rough set theory to enhance student management by identifying at-risk students and enabling personalized interventions. The model combines classification algorithms with rough set-based decision rules to analyze complex student data, including academic performance, behavior patterns, and levels of engagement. The ML layered approach detects patterns and outliers, supporting data-driven decisions to improve student well-being and educational outcomes. Evaluation on the Open University Learning Analytics Dataset (OULAD) demonstrated high accuracy (97.85%) in predicting student outcomes and precision (94.62%) in identifying students needing support. The hybrid approach outperformed conventional methods by approximately 15%, showcasing its transformative potential. This framework effectively monitors student performance and enables customized interventions to meet individual learning needs, fostering a more supportive educational environment.

In this paper we describe an Agile model-driven engineering (MDE) approach, AMDRE, for the re-engineering of legacy systems. The objective is to support the reuse of business-critical functionality from such systems and the porting of legacy code to modernised platforms, together with technical debt reduction to improve the system maintainability and extend its useful life. AMDRE uses a lightweight MDE process which involves the automated abstraction of software systems to UML specifications and the interactive application of refactoring and rearchitecting transformations to remove quality flaws and architectural flaws. We demonstrate the approach on Visual Basic, COBOL and Python legacy codes, including a finance industry case. Significant quality improvements are achieved, and translation accuracy over 80% is demonstrated. In comparison to other MDE re-engineering approaches, AMDRE does not require high MDE skills and should be usable by mainstream software practitioners.

Securing user electronics devices has become a significant concern in the digital period, and a forward-thinking solution covers the fusion of blockchain (BC) technology and deep learning (DL) methods. Blockchain improves device safety by transforming access management, storing credentials on a tamper-resistant ledger, mitigating the risk of unauthorized access and giving a robust defence against malevolent actors. Integrating DL into this framework also raises safety measures, as it permits devices to inspect and regulate to develop attacks distinctly. DL models accurately recognize intricate designs and anomalies, allowing the technique to distinguish and threaten possible attacks in real time. The fusion of BC and DL not only improves the reliability of user electronics but also establishes a dynamic and adaptive safety system, enhancing consumer confidence in the safety of their devices. Therefore, this study presents a BC-Based Access Management with DL Threat Modeling (BCAM-DLTM) technique for securing consumer electronics devices in the IoT ecosystems. The BCAM-DLTM technique mainly follows a two-phase procedure: access management and threat detection. Moreover, BC technology can be applied to the access management of consumer electronics devices. Besides, the BCAM-DLTM technique applies a deep belief networks (DBNs) model for proficiently identifying threats. To enhance the recognition results of the DBN model, the hyperparameter tuning procedure uses the reptile search algorithm (RSA). The experimental outcome study of the BCAM-DLTM approach employs the NSLKDD dataset. The comprehensive results of the BCAM-DLTM approach portrayed a superior accuracy outcome of 99.63% over existing models in terms of distinct metrics.