The use of large datasets for predictive modeling in intelligent databases is frequently hindered by structural deficiencies, most notably missing records. Conventional imputation methods, such as mean or median substitution, distort statistical variance and fail to capture multivariate relationships among features. This paper presents a dynamic, AI-based data quality management framework that autonomously identifies and corrects structural anomalies prior to predictive modeling. Utilizing Classification and Regression Trees (CART), the system intelligently imputes missing categorical and numerical values by learning localized patterns from the observed portion of the dataset. To validate the integrity of the repaired database, the framework was evaluated on a credit risk dataset using a Tri-Ensemble of gradient boosting algorithms (XGBoost, LightGBM, and CatBoost), achieving an Area Under the ROC Curve (AUC) score of 0.9393. These results demonstrate that predictive, tree-based data imputation preserves the statistical distribution of the original data and substantially enhances the accuracy of downstream machine learning tasks. The proposed framework offers a scalable, automated solution for data preprocessing in distributed and intelligent database environments.
Explainable Artificial Intelligence (XAI) is rapidly being integrated into project and program management decision-support tools, driven by the widely held assumption that model transparency improves human decision quality. This assumption has not been subjected to rigorous empirical testing in program management contexts. This paper presents the design, theoretical framework, experimental methodology, and pilot study findings of a controlled behavioural experiment evaluating whether program managers who receive XAI-augmented risk predictions make measurably better decisions than those receiving prediction-only outputs or no AI assistance. Three experimental conditions are defined: unaided judgment, prediction-only AI output, and XAI-augmented AI output with SHAP-derived feature explanations. Four dependent variables are measured: decision accuracy, intervention targeting precision, decision latency, and stakeholder communication quality. A pilot study conducted with 18 practitioners provides preliminary support for the primary hypotheses, with moderate-to-large effect size estimates confirming adequate statistical power for the full study. The framework is grounded in Dual Process Theory, Bounded Rationality, and the Technology Acceptance Model. The study addresses a critical empirical void at the intersection of explainable AI and program governance and is positioned to produce the first empirically validated evidence base on the decision-level value of XAI in IT program risk management.
As organizations transition to hyper-connected, multi-cloud architectures, the mobility of data has far outpaced the technical frameworks meant to govern it. Regulatory mandates GDPR, CCPA, and CPRA articulate purpose limitation and informed consent as foundational principles, yet a persistent gap exists between what these laws require and what distributed systems actually enforce. In cloud-native environments, the metadata that records user consent and allowable use is routinely lost or overwritten as data crosses microservice boundaries, IoT pipelines, and third-party APIs. This article examines the architectural roots of that failure through three documented enforcement actions: the 2026 FTC settlement with General Motors over its Smart Driver telematics program, Tesla's internal camera-data misuse, and the Toyota-Progressive Insurance opt-out breakdown. Each case traces the same systemic flaw, consent does not travel with the data it authorizes. In response, this article proposes a Sticky Governance framework built on cryptographically signed consent tokens, service mesh enforcement, and dynamic resource provisioning, together with a revised audit methodology appropriate for information systems professionals operating in automated, high-velocity data environments. Analysis of three documented enforcement actions reveals a consistent pattern: purpose-scope metadata was absent from inter-organizational API payloads in all cases examined, enabling downstream data misuse that produced a 2026 FTC consent order, internal privacy violations affecting video footage from tens of thousands of enrolled vehicles, and insurance opt-out failures persisting across at least one underwriting cycle. The Sticky Governance Framework operationalizes cryptographic consent propagation and automated purpose-matching at the service-mesh enforcement layer, directly addressing the architectural gap through which governance context is stripped in transit. Policy recommendations include mandatory consent-token schema standards for cross-organizational API ecosystems, governance-metadata preservation clauses in cloud service contracts, and audit frameworks grounded in cryptographic evidence rather than document-review sampling.
This article proposed AI-ADS, an AI-Driven Accelerated Diagnostic System that integrates automated blood analysis, rapid X-ray interpretation, and rapid laboratory orchestration into a single latency-aware architecture. A physiology-structured blood analyzer, a reliability-gated fusion module with uncertainty estimation and budgeted early departures, and a dual-path radiograph encoder with semantic and frequency cues are all part of the proposed solution. In order to translate the speed of diagnostics into actual clinical throughput, AI-ADS employs a digital-twin controller to further model laboratory operations, and constraint-aware scheduling to reduce turnaround-time tails while prioritizing uncertain or critical cases. With an AUROC (macro) of 0.962, an AUPRC (macro) of 0.931, and F1 = 92.2% (Table 1; Fig. 1), the experimental results demonstrate that multimodal fusion improves diagnostic quality in comparison to single-stream baselines. Probability trustworthiness is enhanced by calibration and uncertainty gating, as shown in Table 2 and Figure 2, where ECE = 0.021, Brier = 0.074, and error@high-confidence = 1.9%. Table 3 and Figure 3 show that AI-ADS offers real-time inference with a throughput of 18.4 cases/sec, an early-exit rate of 71%, and a P50/P95 end-to-end latency of 50/120 ms. Table 4 and Figure 4 show that the lab's critical-test compliance increased from 81.2% to 93.5% after using the digital-twin CRL scheduler, and that the median TAT decreased from 49 to 34 minutes. The 90th percentile TAT also decreased from 92 to 62 minutes. Clinical timeliness, diagnostic reliability, and accuracy can all be improved with AI-ADS, as shown here.
This study evaluated the performance of three major SQL injection (SQLi) detection categories—deep learning models, security tools, and structured detection frameworks. Experiments were conducted on a benchmark SQLi dataset derived from publicly available and synthetically augmented SQL traffic, with performance evaluated using accuracy, F1-score, AUC, latency, and false positive rate. Using this dataset containing diverse SQLi variants, the research compared hybrid CNN–LSTM–Autoencoder and Transformer-based models against widely used tools (SQLMap, Acunetix, Microsoft Defender for SQL, CodeScan Labs) and established frameworks (IDE, DIAVA, SQL Shield, ASTF). Deep learning models achieved the highest accuracy (≥0.99), followed by frameworks (0.86–0.96), while tools recorded the lowest detection capability (0.75–0.92). ANOVA results (F = 11.12, p = 0.0013) confirmed statistically significant performance differences. The findings demonstrate the superiority of deep learning—especially hybrid architectures integrating structural, sequential, and latent features—in detecting modern SQLi attacks. This comparative analysis provides empirical evidence supporting the prioritization of adaptive neural models in database security environments.
Sentiment analysis has emerged as a crucial tool for understanding public opinion across various fields, yet studies focusing specifically on political speeches remain limited. This research employs advanced computational techniques to analyze the sentiments expressed in King Abdullah II's speech at the United Nations, which addresses pressing humanitarian concerns amid the ongoing crisis in Gaza. While traditional sentiment analysis typically classifies sentiments into positive, negative, and neutral categories, this study refines this approach by introducing five nuanced classifications: Empathetic, Call to Action, Critique of Power, Reassuring, and Urgent. Utilizing state-of-the-art natural language processing methodologies, we directly analyze the speech text, employing various supervised machine-learning algorithms to assess the effectiveness of these models. Preliminary findings reveal an accuracy rate of approximately 85%, with the Support Vector Machine algorithm demonstrating exceptional performance across all sentiment categories. Notably, The analysis uncovers a predominance of empathetic and critical sentiments, underscoring the King’s deep concern for humanitarian issues and the urgent need for action. This study highlights the classifier performance through metrics such as accuracy, precision, recall, and F1-score and fills a significant gap in the literature by providing patterns into political communication strategies during crises. The findings offer valuable implications for policymakers and community leaders, enabling them to foster civic engagement and inspire citizens to contribute to national development. By enhancing our understanding of the emotional nuances within political discourse, this research serves as a model for future sentiment analysis studies, ultimately contributing to the resilience of communities in the Arab region. Limitations include the single-speaker, single-speech dataset; we discuss directions for broader validation on larger, multi-speaker corpora.
Reviewer Acknowledgements for Computer and Information Science, Vol. 19, No. 1, 2026
This study develops a natural language processing model for sentiment analysis of Queen Rania of Jordan's speeches, focusing on her advocacy for humanitarian causes. Utilizing advanced machine learning and AI techniques, we analyze her communication about refugees and vulnerable populations to identify her advocacy strategies. By integrating technology for nuanced discourse analysis and refining traditional sentiment classifications with contextual information and expert input, we identify key themes and public sentiment trends and quantify emotional responses that may not be apparent from qualitative analysis alone. The research also examines the linguistic features of her speeches, including word choice, tone and rhetorical devices, and considers the socio-political context surrounding her messages to better understand their impact on global humanitarian issues. Given the accessibility of the speeches, a manual classification phase engaged experts to identify key terms and enhance analytic accuracy. The proposed model achieved strong performance (accuracy 87%, precision 82%, recall 80%, F1-score 81%), and after incorporating expert feedback these metrics improved (accuracy 90%, precision 88%, recall 85%, F1-score 86%). Confusion-matrix analysis showed the model reliably distinguished neutral content from positive and negative sentiment, with most misclassifications occurring between neutral and affective categories. Comparison with expert classifications found five discrepancies across empathy, urgency, and hopefulness labels, typically where the model labeled emotionally charged lines as neutral. Findings illustrate the alignment of Queen Rania’s messaging with public sentiment and underscore the role of technology-driven communication in humanitarian advocacy. We identified dominant themes such as compassion and resilience and quantified emotional responses including hope and urgency. Overall, the study aims to illuminate the effectiveness of Queen Rania's communication strategies, demonstrate how sentiment analysis and NLP can deepen understanding of public engagement, and provide actionable guidance to improve humanitarian messaging in the digital age.
A quipu is an Inca device. Made up of knotted strings of varying direction and colour, some of which may be grouped, this device encodes information. Despite ongoing efforts, the meaning of such information remains largely unknown. Unlike existing work, this research does not attempt to decipher the information embedded in a quipu, but instead to apply this traditional Indigenous device to modern software development. A data structure based on the quipu is constructed. Like a quipu, the data structure is hierarchical, partially unsorted, grouped and summed. Interfaces to the data structure in the form of four classes in both C++ and Python are given. A corresponding file format is further proposed. Three applications of the data structure are presented, namely a spreadsheet, a file system and image representation. Practical examples of each of these three applications are presented. This research brings to light the importance of traditional Indigenous knowledge in modern technology.
Digital transformation and digital humanism are reshaping knowledge and interactions across diverse fields and industries. To adapt and fit within the modern digital economy, professional patterns and career pathways require critical and vital capabilities to herald new opportunities and prospects. In light of these dynamics, the research explores the implications of applied artificial intelligence and realities to bridge career prospects in the field of information science, while focusing on digital transformations to foster the development of novel applications. It evaluates essential emerging programs and applications to bridge and advance career projects in the field through the lens of digital transformation; demonstrates how to facilitate effective integration and adoption of applied artificial intelligence technologies and applications; explores factors that hinder these emerging trends and dynamic changes; and formulates a strategic framework to leverage applied artificial intelligence and digital transformation for future opportunities. Research applied content analysis and knowledge from diverse electronic journals, books, online databases, the Internet and the World Wide Web. Research publications and articles were identified and searched using a statistical approach of preferred reporting items for systematic reviews and meta-analyses (PRISMA) strategy and scoping review methodologies. A mixed method research design incorporating quantitative and qualitative approaches was applied to collect and analyze, with concurrent and sequential triangulation used to enhance the validity of the findings. First insights indicate that integration of emerging programs, such as artificial intelligence, machine learning, deep learning, generative artificial intelligence and large language models – reflects transformative technical expertise and strategic innovation in information science, which positions digital transformation as the defining framework to foster interdisciplinary competencies, enhance employability and advance sustainable technological and socio-economic development in the global knowledge economy. Second insights demonstrate that adoption of applied artificial intelligence depends on technological, human and ethical pillars, with digital infrastructure and cloud readiness emerging as the most influential. Third insights highlight multiple interrelated factors that hinder these emerging trends and dynamic changes - inadequate preparedness and training, limited institutional support and resources, resistance to change and lack of awareness of practical AI tools. Fourth insights determine a strategic framework to leverage artificial intelligence and digital transformation for future opportunities from curriculum coordination to technical instruction, as well as AI mentorship and leadership to effectively enhance market competitions and industrial portfolios.
This paper, part of a larger dissertation, challenges the prevailing characterization of humans as the “weakest link” in cybersecurity, a perspective that has led to significant resource misallocation and flawed defensive strategies. Hence, the study empirically investigates the relationship between specific human factors and the severity of security incidents. Employing a sequential explanatory mixed-methods design, this research integrates quantitative analysis of 237 incidents from the VERIS Community Database with qualitative insights from interviews with 12 cybersecurity professionals. The quantitative analysis reveals a critical distinction: human error is associated with a significant reduction in incident severity (odds ratio [OR] = 0.28, p < 0.001), whereas social engineering is linked to a twofold increase in severity (OR = 2.04, p = 0.039). These findings directly challenge the monolithic view of the “human element” and the assumption that initial access vectors reliably predict impact. Qualitative data further illuminate these patterns, indicating that errors are often quickly detected and contained, whereas social engineering facilitates deeper, more persistent intrusions. This study proposes an empirically grounded framework for human-centric incident severity, advocating for a strategic shift from generic awareness training to a dual focus on error-tolerant systems and advanced behavioral detection capabilities. The research offers a refined theoretical lens for understanding human factors in cybersecurity and provides actionable recommendations for optimizing security investments.
Hallucinations in large language models (LLMs) are outputs that are syntactically coherent but factually incorrect or contextually inconsistent. They are persistent obstacles in high-stakes industrial settings such as engineering design, enterprise resource planning, and IoT telemetry platforms. We present and compare five prompt engineering strategies intended to reduce the variance of model outputs and move toward repeatable, grounded results without modifying model weights or creating complex validation models. These methods include: (M1) Iterative Similarity Convergence, (M2) Decomposed Model-Agnostic Prompting, (M3) Single-Task Agent Specialization, (M4) Enhanced Data Registry, and (M5) Domain Glossary Injection. Each method is evaluated against an internal baseline using an LLM-as-Judge framework over 100 repeated runs per method (same fixed task prompt, stochastic decoding at τ= 0.7. Under this evaluation setup, M4 (Enhanced Data Registry) received “Better” verdicts in all 100 trials; M3 and M5 reached 80% and 77% respectively; M1 reached 75%; and M2 was net negative at 34% when compared to single shot prompting with a modern foundation model. We then developed enhanced version 2 (v2) implementations and assessed them on a 10-trial verification batch; M2 recovered from 34% to 80%, the largest gain among the four revised methods. We discuss how these strategies help overcome the non-deterministic nature of LLM results for industrial procedures, even when absolute correctness cannot be guaranteed. We provide pseudocode, verbatim prompts, and batch logs to support independent assessment.
This paper presents a neuro-symbolic framework for detecting cyber compliance violations by integrating deep neural networks with symbolic rule-based reasoning. Traditional machine learning models, while effective in identifying complex patterns, often lack interpretability, limiting their use in regulated domains where explainability is essential. Conversely, symbolic systems offer transparency but are rigid and difficult to scale. Our approach unifies these paradigms by jointly optimizing predictive performance and symbolic rule consistency. Compliance knowledge is encoded as Boolean constraints and incorporated during training as a regularization objective. The model fuses neural embeddings with rule satisfaction signals to improve both accuracy and interpretability. Evaluated on real-world cybersecurity datasets, our method achieves a 99.1% accuracy, 0.96 F1-score, and a reduced false positive rate, outperforming existing baselines. The framework also provides interpretable justifications by identifying violated rules, enhancing trust and auditability. These results demonstrate the feasibility and value of neuro-symbolic systems in scalable, explainable compliance monitoring.
The explosive rise in city dwellers has created a growing scenario of waste management, presenting tremendous environmental, logistical, and health dilemmas to the urban regime. This study focuses on how a more innovative waste management system could be designed and implemented by integrating the IoT devices and Artificial Intelligence (AI) technologies to increase efficiency, transparency, and sustainability of a city's sanitation. The smart bins that would have been introduced in the proposed system would have incorporated IoT and sensors to help understand the current levels of waste, as well as the temperature and humidity of each bin. The transmission of these data is facilitated by low-power, wide-area network technologies such as LoRaWAN and NB-IoT to a centralized waste management system. Such AI algorithms include linear regression, random forest, and A* search, which analyze past and real-time data to refine waste collection schedules, estimate fill levels, and create smart route plans. Moreover, AI models that rely on computer vision can automate the segregation of waste into categories of allowable, organic, and non-recyclable materials with minimal human input. To ensure accountability and track waste throughout its lifecycle, blockchain technology is incorporated, providing tamper-proof records from collection to final disposal. This essay illustrates how the combination of IoT, AI, and blockchain has the potential to reinvent one of the oldest premises of waste management into a data-driven, flexible, and ecologically conscious system that embraces the circular economy concept. Pilot implementation has demonstrated how operational costs are reduced, the recycling rate is increased, and how fuel usage and emissions are significantly decreased.
In this study, the problem of malware proliferation is examined with emphasis on the role of artificial intelligence (AI) in its formation and propagation. The objective of this research is to analyze common malware attacks, their mechanisms, and prevention strategies, drawing upon literature. Methods involve a qualitative review of reported cases and cybersecurity guidelines published between 2010 and 2024. Findings indicate that AI both exacerbates malware threats through adversarial attacks, automated code generation, and phishing automation and offers tools for improved detection and defense. AI-driven anomaly detection, machine learning based intrusion prevention, and adaptive defense systems show promise in mitigating advanced threats. The review also highlights gaps in governance, adversarial machine learning defenses, and protection for IoT and embedded systems. It concludes that addressing malware proliferation requires coherent frameworks, administrative controls, and AI oversight. Future research should prioritize zero-trust architectures, adversarial machine learning defense strategies, supply chain resilience, and governance policies to ensure sustainable and adaptive cybersecurity defenses.
This study concentrates on two primary challenges in the optimization of humanoid robotic arm configurations and the calibration of flexible visual measurement systems, with the goal of improving the motion adaptability and measurement accuracy of robotic systems. In terms of configuration design, we propose a method for screening candidate configurations based on motion flexibility analysis, which incorporates principles from both human anatomy and robotics. By generating performance distribution charts for candidate configurations and comparing them with human arm movement characteristics and workspace parameters, we ultimately identify the most compatible serial robotic arm configuration, establishing a foundation for subsequent motion planning. Regarding calibration optimization, we devise an improved strategy to address the limitations of existing methods. This strategy establishes a joint correction model for hand-eye relationship errors and kinematic parameter deviations, utilizing a linear structured light sensor mounted on the end-effector and fixed reference constraints. Through iterative algorithms that enhance calibration precision, it maximizes the system's potential for high-precision robotic operations. The research offers theoretical and technical support for the synergistic optimization of intelligent control in humanoid robotic arms and high-precision visual measurement systems, demonstrating significant engineering applicability.
In supply chain management, price quotations on machine components can be crucial in determining accurate pricing that may suit a business’s reliability and improve cost control. This process emphasizes price comparisons and negotiations to enable well-informed decisions that do not compromise quality or cost. With the emergence of machine learning and artificial intelligence, companies can leverage these tools to discern reasonable prices and establish the price quotations they can present to suppliers. This study evaluates machine learning models that employ previous material prices and properties such as part family, material grade, dimensions, thread type, and coating to determine the most optimal model. However, the prototype designed within this research expands further by adjusting the hyperparameters of the chosen model to enhance its efficacy. Ultimately, the model analysis determined that Gradient Boosting had a significant predictive accuracy and thus was the best-fit model to forecast pricing on mechanical components.
Personality is an important aspect when deciding group dynamics. Depending on a person’s way of decision-making, or choice of playstyle, this can change a Final Fantasy game’s playthrough. For this study, the playstyles of Final Fantasy X in Japan and the United States were surveyed. This research involved using Hofstede’s cultural dimensions theoretical framework combined with Likert scale player responses. This research project aimed to show that individual player decisions aligned with the choices expected and predicted by the Hofstede’s cultural dimensions framework. The project focused on five of the Hofstede dimensions: individualism/collectivism, masculinity/femininity, time orientation, uncertainty avoidance, and restraint/ indulgence. The data collected validated this hypothesis in the majority of the dimensions studied.
The neighborhood-based bridge node centrality (NBNC) tuple has been proposed in the literature to rank nodes for the extent they could serve as a bridge node. The NBNC tuple of a node v has three entries: (# components in NGv, 1-algebraic connectivity ratio of NGv and degree of node v), where NGv is the neighborhood graph of node v. The research presented in this paper conducts principal component analysis on dataset comprising of NBNC tuples of all the nodes and computes a weighted PC_NBNC score based on the entries for the nodes in the dominating principal components (variances ≥ 1.0). The proposed model is to classify nodes as core (or peripheral) if their weighted PC_NBNC score is ≥ 0.0 (or < 0). The study measures the fractions of core-core, core-peripheral and peripheral-peripheral links and the fractions of core and peripheral nodes and uses these measures to classify a real-world network as either core-heavy or peripheral-heavy. Accordingly, 48 of the 80 real-world networks are classified as core-heavy (observed to be dominated by core nodes and core-core links) and the remaining 32 networks are classified as peripheral-heavy (observed to be dominated by peripheral nodes).
Reviewer Acknowledgements for Computer and Information Science, Vol. 18, No. 1, 2025
