This systematic literature review (SLR) investigates interaction strategies in augmented reality (AR)-situated visualizations that support decision-making tasks. Despite its growing relevance, a gap remains in understanding how users interact with these systems to perform decision-related tasks. This review aims to examine how various interaction modalities contribute to effective user engagement in decision-making contexts. A total of 23 peer-reviewed studies, published between 2016 and 2024, were analyzed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure methodological rigor and conceptual alignment. The review focuses on device types, manipulation methods, and interaction modalities. Findings indicate that head-mounted displays (HMDs) are commonly used in immersive settings, while hand-held devices (HHDs) are preferred for their portability and affordability. Touchscreen interaction dominates among HHDs, enabling direct manipulation, while gesture and voice commands are more prominent in HMD-based systems. A conceptual mapping aligns these modalities with manipulation methods such as selection, navigation, and filtering based on Brehmer and Munzner’s typology. This review also highlights key challenges, including limited collaborative support, usability concerns, and underrepresentation of domains such as industrial training and public events. This work provides a structured foundation for designing AR-situated visualization systems that better support decision-making across diverse application contexts.

Tuberculosis (TB) remains a critical global health challenge, particularly in resource-constrained settings where timely and accurate diagnosis is essential for effective disease management and control. Traditional diagnostic methods, such as Ziehl-Neelsen (ZN)-stained sputum microscopy, are widely employed for detecting Mycobacterium tuberculosis; however, these techniques are inherently subjective and prone to variability due to their reliance on manual interpretation. In response, an increasing body of research has applied deep learning (DL)-based approaches to automate TB detection from microscopy images. This systematic review synthesizes findings from 67 studies that have explored various machine-learning techniques for TB diagnosis using ZN-stained images. A structured literature search was conducted across multiple scientific databases, including PubMed, IEEE Xplore, Scopus, and ScienceDirect. Studies were selected based on their focus on DL applications for TB detection using ZN-stained images. The reviewed methodologies encompass various stages, including image preprocessing, feature extraction, classification strategies, and performance evaluation metrics. Our review reveals that DL models, particularly those employing automated feature extraction and classification, are predominantly used, with some studies reporting accuracies of up to 100%. This review provides a comprehensive overview of state-of-the-art methodologies, including image preprocessing, feature extraction, classification strategies, and performance evaluation metrics. Notably, the evidence indicates that convolutional neural network (CNN)-based approaches offer the highest promise due to their robust ability to detect subtle features in stained images. Consequently, future research will focus on developing and optimizing CNN-based models to further enhance TB detection, ultimately improving diagnostic outcomes and supporting more effective TB control strategies.

Predicting sports performance has become a central focus in sports analytics, driven by the increasing availability of data and the growing recognition of its potential impact on decision-making in the sports sector. Time series analysis and real-time prediction of athletic performance involve forecasting an athlete’s performance over time, allowing coaches and sports scientists to refine training programs, manage workload, and make informed strategic decisions. This study thoroughly examines time series prediction and real-time prediction in sports, as well as the artificial intelligence (AI) techniques employed by prior researchers. The review is conducted with precision, strictly following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) criteria. This article examines time series prediction and real-time prediction methodologies that utilize machine learning (ML) and deep learning (DL) approaches, spanning the period from 2020 to 2025. This article covers the range of AI methodologies from the most basic to the most advanced models. A detailed assessment of ML and DL methodologies, grounded in prior research findings, is presented. All approaches examined in this paper significantly influence the primary future study, which focuses on the hybrid long short-term memory (LSTM) peephole integration with gated recurrent unit (GRU) for use in track cycling sports, the principal objective of the research. This research is consistent with the United Nations’ Sustainable Development Goals.

This study aims to develop an agricultural land recommendation system by integrating the Internet of Things (IoT) and machine learning (ML). IoT devices, including the JXBS-3001 soil sensor and Raspberry Pi Pico RP2040, collect real-time soil data, which is analyzed using the decision tree (DT) algorithm. The DT algorithm is chosen for its simplicity, efficiency, and interpretability over random forest (RF) and k-nearest neighbors (k-NN). It provides structured decision-making, faster training, and better handling of numerical data for parameters such as soil pH, nutrient content (NPK), moisture levels, and temperature. The findings show that the system provides accurate crop recommendations, helping farmers make informed decisions. The integration of IoT and ML enhances land assessment and optimizes agricultural productivity. Future improvements could include weather analysis and plant disease detection to further support decision-making.

The increasing demand for sustainable agricultural practices has led to the exploration of alternative composting methods, including the use of black soldier fly larvae (BSFL) for managing organic waste. This study compares the physical, chemical, nutrient, and nutrient loss properties of frass produced through two treatments at specified moisture contents of 65 and 80%, respectively (BSFL-65 and BSFL-80). The physical properties, including yield, water holding capacity, bulk density, and moisture content, were analyzed, revealing no significant differences between the treatments for yield, water holding capacity, and bulk density. However, moisture content was significantly higher in BSFL-80, indicating a distinct impact of treatment on water retention. In terms of chemical properties, the pH of BSFL-80 was significantly higher than that of BSFL-65, while electrical conductivity showed a significant difference at the borderline. Total dissolved solids were significantly higher in BSFL-80, whereas total volatile solids were significantly higher in BSFL-65. Nutrient content analysis revealed no significant differences in carbon (C), hydrogen (H), sulfur (S), nitrogen (N), phosphorus (P), potassium (K), and other elements, except for ammonia nitrogen, which was slightly higher in BSFL-80. Nutrient loss assessments showed no significant differences between the treatments for C, H, S, N, P, K, calcium, magnesium, copper, or zinc. These findings suggest that while certain chemical and physical properties, such as moisture content and pH, are significantly influenced by the treatment type, the overall nutrient composition and nutrient loss were largely unaffected by the treatment variations. This study provides valuable insights into the comparative efficacy of BSFL-based frass treatments, offering potential implications for sustainable agricultural practices and waste management.

The effectiveness and efficiency of chemical spray applications on cropland are strongly influenced by the size of the nozzle. This research study focuses on examining the spray characteristics associated with two different nozzle sizes using bio-fertilizer. The experimentation took place on a 13.85-ha paddy field. To collect spray deposits, water-sensitive papers (WSPs) were strategically placed in a setup consisting of 39 WSPs arranged at distances of 0.5 m apart in three rows spaced 1 m apart. An agricultural spraying drone was deployed at a speed of 3 m/s and a height of 2 m to conduct the tests. Parametric data, essential for both descriptive and inferential statistical analyses, were generated using the DepositScan software. The results of the characteristic test indicated that approximately 50% of the collected droplet sizes were below 89.11±20.59 and 239±78.44 µm for the fine and coarse nozzle sizes, respectively. The coefficients of variation were calculated as 0.27 for the very fine nozzle and 0.85 for the coarse nozzle, indicating varying levels of droplet size uniformity. The t-test analysis revealed a significant difference (P ≤ 0.05) between the two nozzle sizes for all investigated parameters, except for the drift collected within a 30-m distance from the target. Specifically, the fine nozzle type produced very fine droplets with superior spray uniformity, while the coarse nozzle types generated moderate to large droplet sizes. This highlights the critical role of selecting the appropriate nozzle in determining the overall quality of spray applications.

Due to the exponential growth of online information, the ability to efficiently extract key content and target information without requiring extensive reading is becoming increasingly important for readers. This paper investigates the construction of neural extractive summarization systems by framing the task as a semantic text matching problem. The proposed approach, named MatchDocSum, aligns the source document with potential summaries within a semantic space, leveraging pretrained language model contextual representations to enhance the understanding of their interconnectedness. The goal is to address the limitations of conventional methods, which often struggle with capturing intricate semantic relationships and producing coherent summaries. Hence, this study proposes an enhanced document summary matching framework to investigate three main aspects that affect the outcome of a good summary: document pruning, text embedding, and similarity matching measure within the framework. MatchDocSum was evaluated on the Cable News Network (CNN)/DailyMail dataset, showing competitive results against several baselines, including LEAD and bidirectional encoder representations from transformers (BERT) for extractive summarization (BERTSUM). The results demonstrate that our approach performs better than baseline models in some aspects, achieving Recall-oriented Under for Gisting Evaluation (ROUGE)-1 scores of 43.50, ROUGE-2 scores of 20.45, and ROUGE-L scores of 40.75.

Power cables play a crucial role in modern power transmission and distribution networks, serving as vital infrastructure for national grids and power utility companies. However, their high installation, operation, and maintenance costs pose significant challenges, especially as global electricity demand continues to rise. Optimizing power cable technology is essential to ensuring reliable, cost-effective, and sustainable energy systems. These challenges align with the United Nations' Sustainable Development Goals (SDGs), particularly in providing access to affordable and clean energy. While extensive research has been conducted on the theoretical and practical aspects of power cables, a noticeable gap remains in bibliometric and scientometric analyses in this field. Understanding research trends, collaboration networks, and emerging technologies is essential for guiding future developments. This study presents the first comprehensive scientometric analysis of power cable research, addressing the gap by systematically evaluating the field's evolution. Using data from the Web of Science (WoS) database and CiteSpace for visualization, the study examines publication trends, key thematic areas, and the global impact of scholarly contributions. The findings reveal the leading researchers, the most influential publications, the volume of research output by country, and the extent of international collaboration. Additionally, the analysis highlights emerging trends in power cable technology, including advancements in materials, diagnostics, and the integration of smart grids. By mapping the research landscape, this study provides valuable insights for academics, industry professionals, and policymakers to foster collaboration, drive innovation, and enhance the development of sustainable power infrastructure.

A hybrid laminated composite is a laminate formed by integrating composite layers from various fibre types to achieve optimal properties. Nonetheless, considerable knowledge remains to be acquired regarding the vibration behaviour associated with the hybridisation of composite laminates. In structural design, natural frequency is a critical factor for preventing resonance, which may result in significant structural failure. This study aims to assess the inherent natural frequency response of hybrid composite laminates under free vibration, influenced by varying plate thicknesses, layer fractions, and orientation angles. Finite element models were developed using a commercial finite element software, ANSYS, to precisely characterise the natural frequencies of hybrid composite laminates under free vibration. The design of experiments was employed to identify 17 case study runs and to assess significant factors, with a comprehensive examination of each factor's impact on natural frequencies conducted through modal analysis. Given the limited dataset, employing techniques such as cross-validation with response surface methodology (RSM) and artificial neural networks (ANN) enhances the reliability of performance assessment for the model. Optimisation was conducted utilising RSM via analysis of variance, while ANN serves as a tool to ascertain data accuracy. The accuracy and robustness of the models are corroborated by a comparison of predictions from finite element analysis and RSM, demonstrating a strong correlation with percentage errors of 16 and 10% for ANN, respectively.

This study presents a comparative analysis of nonlinear regression models integrated with feature selection for predicting the cleanliness factor (CF) in coal-fired utilities. The models evaluated are regression trees (RT), support vector regression (SVR), ensembles of trees, and artificial neural networks (ANN). Different boiler designs introduce various operational parameters that influence cleanliness, making it more challenging to predict real-time data accurately. To enhance the model’s predictive accuracy, the minimum redundancy maximum relevance (MRMR) feature selection technique was integrated, facilitating improved model performance by selecting the best subsets of variables. Model performance was assessed accordingly, where the number of selected features varies between 138 and 10. The results indicate that a combination of bagged trees and MRMR with 10 features achieved R² values of 0.973 for the training set and 0.976 for the test set, with a mean squared error (MSE) of 0.001 for both datasets. Compared to SVR and ANN, bagged trees consistently demonstrated superior predictive accuracy with reduced computational complexity. These findings confirm that ensemble-based models, particularly bagged trees with MRMR, provide the most accurate and computationally efficient approach for CF prediction. An accurate CF prediction creates more reliable information for a data-driven approach that solves the soot-blowing operational system. The system has the risk of either underblowing or overblowing steam during boiler cleaning. This risk, if not properly handled, may lead to more severe ash fouling and slagging issues, such as emergency shutdowns, metal corrosion, and declining heat transfer efficiency in coal-fired utilities. Overall, improving real-time boiler monitoring minimizes steam waste during soot-blowing operations.