In recent years, the integration of Artificial Intelligence (AI) techniques into Functional Size Measurement (FSM) has gained increasing attention within the software engineering community. AI-based FSM methods aim to overcome the limitations of traditional approaches, namely, reliance on expert judgment and structured inputs, by enabling automation and scalability. This study presents a systematic literature review of research efforts that apply AI techniques to FSM. The review analyses 27 primary studies published between 2008 and 2025, sourced from major digital libraries. Various aspects are investigated, including publication trends, AI technique categories, functional sizing methods, and research types. The findings show a predominance of COSMIC as the preferred FSM method and highlight NLP, and transformer-based models as the most commonly employed AI techniques. While hybrid models combining rule-based systems with deep learning show promise, the field remains fragmented and underexplored. This review offers the first consolidated overview of AI-driven FSM research, identifying key trends, challenges, and directions for future investigation.

Recent advancements in artificial intelligence (AI) have strengthened the interaction between sports and digital technologies. However, unlike widely studied sports such as football and basketball, handball has received limited attention from the scientific community, despite its fast-paced nature and strategic importance. This study focuses on object detection in handball and targets key entities, such as players, referees, goalkeepers, and the ball. A comprehensive dataset was created through a collaborative annotation process, consisting of annotated images extracted from real handball games. The YOLOv8 model was then trained and evaluated on this dataset to assess its effectiveness in entity recognition. The proposed approach achieved an object detection accuracy of 86.8% on a relatively small held-out test set, providing an indicative first benchmark for the application of state-of-the-art machine learning models to handball. To the best of our knowledge, the dataset generated in this study is the first comprehensive collection of annotated handball images, providing a valuable resource for further research. By bridging sports analytics and computer vision, this study contributes to the advancement of performance assessment in handball. These exploratory results suggest potential directions for future real-time systems and practical applications, such as improved understanding of player performance, team dynamics, and strategic decision-making.

Efficiency and sustainability are critical considerations in the development and deployment of machine learning (ML) applications. Among the factors influencing sustainability, resource leaks in ML code can introduce hidden inefficiencies that elevate energy consumption and CO2 emissions. Despite this, empirical evidence quantifying their environmental impact remains limited. This emerging results paper presents an initial empirical investigation of two common resource-leak smells, namely Improper Model Reuse (IMR) and Unreleased Tensor References (UTR), and their impact on energy consumption and CO2 emissions in TensorFlow and Keras workloads. Controlled experiments were conducted for each smell by executing identical training tasks while comparing against a smell-free baseline. Our preliminary results show that both smells consistently increase estimated electricity usage and carbon emissions. IMR and UTR increased electricity consumption by approximately 32

Rapid growth in quantum computing has created an urgent need to adapt traditional software engineering methodologies. Among these methodological challenges, software size measurement plays a central role in project estimation, quality assurance, and software engineering management. However, there is no consolidated understanding of how it should be applied to quantum software systems. This paper presents the first systematic literature review on quantum software size measurement, synthesizing existing approaches, and identifying key research gaps. An analysis of 16 studies shows that size measures span multiple abstraction levels, with most research focusing on adapting classical metrics, such as Lines of Code (LOC) and functional size, to quantum software systems. LOC continues to be used for sizing quantum code, while COSMIC Function Points (ISO 19761) have been extended to capture data movements in quantum components, enabling their application at the requirement, code, and circuit levels. At the specification and architectural design levels, the proposed measures include counts of quantum classes and methods as well as lines of architectural specifications. At the circuit level, quantum-specific constructs, such as qubits, gates, and measurements, are also incorporated. Additionally, software structure metrics, including cyclomatic complexity and Halstead metrics, are used to measure quantum software at both the code and design levels. Research output has grown since 2024, and is driven by academic groups, with minimal industrial participation. However, most metrics remain conceptual, lack empirical validation, and are not integrated into development life cycles, highlighting the immaturity and fragmentation of current proposals and the need for standardized frameworks.

Traceability mechanisms ensure transparency and accountability in distributed systems, where autonomous entities interact concurrently with shared resources without complete knowledge of the global state. Most existing approaches are descriptive: they record interactions after execution without constraining them according to the predefined usage rules. This limitation is important when shared resources must be accessed in compliance with global constraints under concurrent executions. We introduce directed traceability as a protocol-level mechanism that governs interaction admission according to usage constraints and produces a compliant execution trace by construction. Unlike descriptive mechanisms or locally-validated approaches such as smart contracts, which enforce only local constraints, directed traceability achieves compliance through execution-time validation against global state. We formalize the problem through a multi-agent model and establish three theoretical properties: 1) directed mechanisms guarantee zero violation rate; 2) descriptive and locally-validated mechanisms exhibit structural violations under concurrency and global constraints; and 3) directed traceability induces a bounded latency overhead of O(k & sdot;(n+m)) , where k is the number of constraints, n the number of agents, and m the number of resources. Numerical evaluation through a distributed container-based implementation yields violation rates of 28%-35% (descriptive), 5%-10% (locally validated), and 0% (directed); a rise to 10% violations for local validation under global constraints; and a latency factor of approximately 100 & times; for the directed mechanism under the tested container-based deployment. Compliance follows from protocol design rather than post-execution audit.

Coordinating autonomous and heterogeneous artificial intelligence (AI) systems is difficult because their capabilities, domain specializations, interfaces, and output structures vary widely and rapidly evolve. The purpose of this study is to address these issues through the design of an AI of the AI (AoA) ecosystem as a hierarchical system-of-systems framework that orchestrates and challenges multiple independent AI agents while preserving their operational autonomy. In this approach, the AoA integrates three interconnected layers: (i) a configuration and longitudinal performance-tracking layer that maintains operational parameters, version histories, and domain-specific performance profiles; (ii) a moderated collaboration and evaluation layer that enables indirect coordination through standardized response schemas, structured competition, and cross-agent benchmarking; and (iii) a context-dependent authority and data-source evaluation layer that weights AI outputs based on source credibility, reference quality, and domain relevance, supported by federated ontologies for semantic alignment and conflict reconciliation. In conclusion, this ecosystem also includes multi-stage alert mechanisms that detect drift, inconsistencies, conflicts, and emergent patterns, enabling continuous self-assessment, adaptive governance, and iterative ranking updates. In terms of practical implications, this AoA is designed to interoperate with existing information technology (IT) infrastructure and is additionally forward-compatible with emerging quantum computing platforms, providing a scalable foundation for orchestrating and governing heterogeneous AI ecosystems.

E-government portals are crucial for public service delivery, yet ensuring their accessibility for all citizens, including individuals with disabilities, remains a pressing challenge. This study addresses this issue by proposing an accessibility-based E-Government Portals Best Practices Model (E-GPBPM). The research begins by establishing a comparative analysis of the three most cited accessibility guidelines in the literature: WCAG, Section 508, and E-MAG. As WCAG 2.0 is the most established and comprehensive standard, it was selected for further analysis. Subsequently, a mapping study was conducted to investigate the extent to which the existing E-GPBPM covers the WCAG 2.0 accessibility guideline. This involved a detailed mapping of the model's best practices against the WCAG 2.0 success criteria. The results show that the WCAG criteria comprehensively align with the E-GPBPM's specific goals and practices within the web content category. Finally, an accessibility-based version of the EGPBPM is proposed to foster inclusive access to digital public services.

Accurate sizing of software systems is essential for effective project management and resource allocation. Hybrid classical-quantum software systems consist of a quantum component, which executes quantum algorithms and incorporates error correction mechanisms (overhead)-a critical aspect in the Noisy Intermediate-Scale Quantum (NISQ) era-and a classical component, which controls quantum operations and manages error propagation. This paper explores functional sizing for hybrid classical-quantum software. It highlights the challenges arising from the interaction between the classical and quantum components, particularly with respect to the abstraction layers and granularity levels. A key principle emphasized is that functional sizing should not extend across multiple abstraction layers within a single measurement scope. While quantum software typically operates at a low-level implementation-specific layer, classical software generally functions at a higher level of abstraction. This divergence emphasizes the need to clarify how functional size measurement should be applied using set theory to account for requirements specified at different abstraction layers in hybrid computing environments.

The rapid evolution of quantum computing has led to growing interest in the development of systematic approaches to assess and manage quantum software. Among these, functional size measurement (FSM) offers a promising pathway for establishing metrics that can support project estimation, benchmarking, and quality assessment. This paper provides an overview of emerging efforts in quantum software functional size measurement with a focus on key concepts, challenges, and motivations. First, we examine the foundational principles of quantum computing in relation to the existing FSM methods, identifying both compatibilities and unique constraints. Building on this, we discuss the challenges that arise when applying FSM to quantum software, including the hardware-coupled nature of functionality, evolving semantics of data and measurement abstractions, gaps in tooling and standardization, and evolving role of FSM across the Noisy Intermediate-Scale Quantum (NISQ) and future faulttolerant eras. We then outline community-driven initiatives, such as the COSMIC Quantum Software Taskforce and the Fall 2024 Workshop, which highlight the increasing demand for structured measurement practices. By articulating these issues, this paper aims to present initial research efforts for quantum software functional size measurement and stimulate further exploration of measurement approaches that are both theoretically grounded and practically applicable in quantum software engineering.

The patient problem list is a key component of an electronic health record (EHR) and must be accurate and accessible for all professionals involved in patient care. Unfortunately, such a list is mostly found in an unstructured text format, is not easily sharable across digital health systems, and lacks semantic interoperability. Natural language processing (NLP) techniques are widely used for clinical concept extraction, particularly for English text. However, in the Canadian context, the clinical notes in a patient problem list can also be found in French. This research presents a framework based on Fast Healthcare Interoperability Resources (FHIR) consisting of an NLP clinical pipeline and a rule-based approach to converting the textual patient problem list, including notes regarding allergies, into an FHIR model. The proposed approach considers concept modifiers to map to the International Patient Summary (IPS) FHIR model element. The main contributions of this research include the early detection of FHIR resources from unstructured data written in the French language and the design of a rule-based algorithm to identify and map extracted data to the appropriate FHIR resource attributes using an annotator. A primary evaluation of the resource tag which uses the rule-based method demonstrates the feasibility of the proposed model to facilitate semantic interoperability. The assessment was conducted using the French FRASIMED corpora.

This study presents ensemble machine learning (ML) models for predicting residential energy consumption in South Africa. By combining the best features of individual ML models, ensemble models reduce the drawbacks of each model and improve prediction accuracy. We present four ensemble models: ensemble by averaging (EA), ensemble by stacking each estimator (ESE), ensemble by boosting (EB), and ensemble by voting estimator (EVE). These models are built on top of Random Forest (RF) and Decision Tree (DT). These base predictor models leverage historical energy consumption patterns to capture temporal intricacies, including seasonal variations and rolling averages. In addition, we employed feature engineering methodologies to further enhance their predictive abilities. The accuracy of each ensemble model was evaluated by assessing various performance indicators, including the mean squared error (MSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination R2. Overall, the findings illustrate the efficiency of ensemble learning models in providing accurate predictions for residential energy consumption. This study provides valuable insights for researchers and practitioners in predicting energy consumption in residential buildings and the benefits of using ensemble learning models in the building and energy research domains.

To fully leverage Google Analytics and derive actionable insights, web analytics practitioners must go beyond standard implementation and customize the setup for specific functional requirements, which involves additional web development efforts. Previous studies have not provided solutions for estimating web analytics development efforts, and practitioners must rely on ad hoc practices for time and budget estimation. This study presents a COSMIC-based measurement framework to measure the functional size of Google Analytics implementations, including two examples. Next, a set of 50 web analytics projects were sized in COSMIC Function Points and used as inputs to various machine learning (ML) effort estimation models. A comparison of predicted effort values with actual values indicated that Linear Regression, Extra Trees, and Random Forest ML models performed well in terms of low Root Mean Square Error (RMSE), high Testing Accuracy, and strong Standard Accuracy (SA) scores. These results demonstrate the feasibility of applying functional size for web analytics and its usefulness in predicting web analytics project efforts. This study contributes to enhancing rigor in web analytics project management, thereby enabling more effective resource planning and allocation.

Much of the existing ML research focuses on model performance metrics, leaving limited attention to the long-term sustainability and resource efficiency of ML applications. While high performance is essential, ensuring efficient resource management is equally critical for robust deployment. This study addresses this gap by systematically identifying code smells that lead to resource leaks in ML applications. We conducted an empirical investigation of developer discussions and real-world code snippets from PyTorch, TensorFlow, and Keras. The analysis identified 30 PyTorch-related smells and 16 TensorFlow/Keras smells linked to resource leaks. These smells were categorized in two ways: (1) based on their root causes, and (2) as general ML smells with framework-specific characteristics. For each smell, we derived at least one best practice, resulting in 50 recommended coding patterns aimed at reducing resource leakage and improving efficiency. To ensure the validity of our findings, we employed a three-phase validation process involving independent analysis by three authors followed by consensus discussions. This is the first comprehensive study to examine resource-leak-inducing code smells across major ML frameworks and to present actionable best practices for mitigating them. The contributions support developers in building more efficient and sustainable ML applications and offer a structured view of the underlying causes of resource leaks.

Effective energy consumption forecasting in smart homes is vital for optimising resource usage and integrating renewable energy sources. Current research is often hindered by factors such as limited or low quality data, the challenge of identifying suitable prediction models, the variability of consumption patterns, and the scalability limitations of these models. In this study, we examine the scalability and performance of three advanced machine learning frameworks: Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and Artificial Neural Networks (ANNs) for short-term energy consumption forecasting across multiple dataset sizes. Specifically, it used four generated datasets representing 20, 50, 100, and 200 smart homes, each covering 365 days of energy consumption data. We assess how well each model adapts to growing dataset sizes by measuring root mean square error (RMSE), mean absolute error (MAE), coefficient of determination ( R^2 ), training time, and inference speed. We examine the models’ ability to generalize from smaller to larger datasets and their suitability to capture diverse consumption patterns in different household datasets. In addition, we assess the resource and time efficiency of each model. Our findings show that ANN models deliver reliable and precise predictions for energy consumption, making them particularly suitable for both residential-scale and city-wide smart energy management.

Embedded systems omnipresent in everyday life and industry are mainly composed of hardware and software that must comply with a number of standards and regulations. However, there is no consensus on the quality characteristics and subcharacteristics of embedded software. This article presents the steps for modeling an operational quality model for embedded software aligned with the ISO 25000 series of quality models for traditional computer systems. From a literature review composed of 40 studies on quality modeling for embedded systems and software, 85 of the most frequent quality characteristics and subcharacteristics were first identified, including a subset of 16 referenced or cited in at least 25% of the literature. Next, the design of a quality model for embedded software aligned with the ISO 25000 series was proposed with 13 characteristics and 27 subcharacteristics. The operational aspect of this quality model for embedded software is addressed next through a set of measures and measurement functions from ISO 25000 to aggregate the results of the quantification of the characteristics and subcharacteristics. A survey involving 25 embedded software specialists is presented next to gauge, using Fleiss's Kappa criteria, their agreement with the proposed quality model. Furthermore, the computed importance weights derived from the survey participants' individual opinions were compared with those derived from an analysis of 40 embedded software studies, bolstering the credibility of the model. The results of this study suggest that the proposed quality model can serve as a framework for evaluating and understanding the quality characteristics across diverse expertise levels. Furthermore, the convergence between the survey and the literature strengthens the model's credibility by anchoring it in both established literature and practitioners' agreements.

Context: Software maintenance is time-consuming and requires significant effort for bug resolution and various types of software enhancement. Estimating software maintenance effort is challenging for open source software (OSS) without historical data about direct effort expressed in terms of man-days, compared to proprietary software for which this data about effort is available. Therefore, maintenance efforts in the OSS context can only be estimated indirectly through other features, such as OSS bug reports, and other approaches, such as bug resolution prediction models using a number of machine learning (ML) techniques. Although these bug reports are at times large in size, they need to be preprocessed before they can be used. In this context, instance selection (IS) has been presented in the literature as a way of reducing the size of datasets by selecting a subset of instances. Additionally, ML techniques often require fine-tuning of numerous parameters to achieve optimal predictions. This is typically done using tuning parameter (TP) methods. Objective: The empirical study reported here investigated the impact of TP methods together with instance selection algorithms (ISAs) on the performance of bug resolution prediction ML classifiers on five datasets: Eclipse JDT, Eclipse Platform, KDE, LibreOffice, and Apache. Method: To this end, a set of 480 ML classifiers are built using 60 datasets including the five original ones, 15 reduced datasets using Edited Nearest Neighbor (ENN), Repeated Edited Nearest Neighbor (RENN), and all-k Nearest Neighbor (AllkNN) single ISAs, and 40 reduced datasets using Bagging, Random Feature Subsets, and Voting ensemble ISAs, together with four ML techniques (k Nearest Neighbor (kNN), Support Vector Machine (SVM), Voted Perceptron (VP), and Random Tree (RT) using Grid Search (GS) and Default Parameter (DP) configurations. Furthermore, the classifiers were evaluated using Accuracy, Precision, and Recall performance criteria, in addition to the ten-fold cross-validation method. Next, these classifiers are compared to determine how parameter tuning and IS can enhance bug resolution prediction performance. Conclusion: The findings revealed that (1) using GS with single ISAs enhanced the performance of the built ML classifiers, (2) using GS with homogeneous and heterogeneous ensemble ISAs enhanced the performance of the built ML classifiers, and (3) associating GS and SVM with RENN (either used as a single ISA or implemented as a base algorithm for ensemble ISAs) gave the best performance.

In the rapidly evolving landscape of modern education, Science, Technology, Engineering and Mathematics (STEM) disciplines stand out for their role in equipping students with the skills necessary to address complex real-world problems. Despite the critical importance of STEM careers, disparities persist, particularly among young women and students from public schools. This motivated the creation of an IoT bootcamp - referred to as PyTime IoT, a 2-day course aimed at high school students. This bootcamp integrates hands-on activities in Python programming and IoT systems that demonstrate the tangible application of theoretical STEM concepts through real-world scenarios. Our approach enhances participants' understanding of IoT and programming and also serves as a stepping-stone for high school students contemplating STEM careers. Preliminary results from the first two bootcamp editions indicate a positive shift in participants' skills and a marked increase in STEM career interest, suggesting that targeted bootcamps like PyTime IoT are effective in bridging the educational gap and inspiring the next generation of STEM professionals.

Background The increasing use of electronic health records and the Internet of Things has led to interoperability issues at different levels (structural and semantic). Standards are important not only for successfully exchanging data but also for appropriately interpreting them (semantic interoperability). Thus, to facilitate the semantic interoperability of data exchanged in health care, considerable resources have been deployed to improve the quality of shared clinical data by structuring and mapping them to the Fast Healthcare Interoperability Resources (FHIR) standard. Objective The aims of this study are 2-fold: to inventory the studies on FHIR semantic interoperability resources and terminologies and to identify and classify the approaches and contributions proposed in these studies. Methods A systematic mapping review (SMR) was conducted using 10 electronic databases as sources of information for inventory and review studies published during 2012 to 2022 on the development and improvement of semantic interoperability using the FHIR standard. Results A total of 70 FHIR studies were selected and analyzed to identify FHIR resource types and terminologies from a semantic perspective. The proposed semantic approaches were classified into 6 categories, namely mapping (31/126, 24.6%), terminology services (18/126, 14.3%), resource description framework or web ontology language–based proposals (24/126, 19%), annotation proposals (18/126, 14.3%), machine learning (ML) and natural language processing (NLP) proposals (20/126, 15.9%), and ontology-based proposals (15/126, 11.9%). From 2012 to 2022, there has been continued research in 6 categories of approaches as well as in new and emerging annotations and ML and NLP proposals. This SMR also classifies the contributions of the selected studies into 5 categories: framework or architecture proposals, model proposals, technique proposals, comparison services, and tool proposals. The most frequent type of contribution is the proposal of a framework or architecture to enable semantic interoperability. Conclusions This SMR provides a classification of the different solutions proposed to address semantic interoperability using FHIR at different levels: collecting, extracting and annotating data, modeling electronic health record data from legacy systems, and applying transformation and mapping to FHIR models and terminologies. The use of ML and NLP for unstructured data is promising and has been applied to specific use case scenarios. In addition, terminology services are needed to accelerate their use and adoption; furthermore, techniques and tools to automate annotation and ontology comparison should help reduce human interaction.

Effective statistical methodologies play a pivotal role in ensuring the validity and reliability of scientific findings. This study unveils critical gaps in statistical methodologies in metaheuristics research. This is done using a quantitative evaluation framework based on standard methodologies employed in the mature scientific fields of medicine, pharmacology, and psychology. A random sample of 70 peer-reviewed metaheuristics articles published between 2018 and 202 3was analyzed. Our analysis reveals that authors favor descriptive statistics (79%, 95% CI [69.8, 88.2]) over rigorous null-hypothesis statistical testing and rarely do (< 5%) report confidence intervals or effect sizes. Of the 43%, (95% CI [38.0, 48.0]) that used statistical testing, a strong majority (> 80%) do not discuss underlying assumptions, control for family-wise Type 1 error rate or provide exact test statistics. These findings reveal a fundamental disparity between current practices and established standards as well as the pressing need for improving statistical rigor in metaheuristics research.