-
Research Article
The Compression Sensing Reconstruction for the 1-d Signal Based on Non-local Full Connection Layer
Issue:
Volume 11, Issue 1, June 2026
Pages:
1-7
Received:
11 December 2025
Accepted:
29 December 2025
Published:
20 February 2026
Abstract: The compression sensing reconstruction for the 1-d signal can contribute to the communication of autonomous driving, intelligent robots, and fire exploration robots. To address the issue that fully connected layers in the LISTA method lack the ability to extract non-local features, this paper primarily designs a non-local fully connection layer and proposes a novel compressed sensing reconstruction method for audio signals. This paper designs a compression sensing reconstruction method for the 1-d signal. To reconstruct 1-d signal, the deep learning method LISTA is used. Then, the linear full connection layer in LISTA is improved by combining the output of three full connection layer to capture the non-local information. The computing regions of improved non-local full connection layer contain: 1) the full connection before; 2) the current full connection; and 3) the full connection after. Experimental results show the reconstruction results of LISTA and LISTA_nf are both close to the real signal. The MSE of LISTA_nf is reduced by 0.1 than the MSE of ISTA under the same experimental settings. The non-local full connection layer in the LISTA_nf consumes longer computing time. The LISTA_nf increase the computing time by 0.07s than the computing time of the ISTA. Experimental results show the effectiveness of the proposed method.
Abstract: The compression sensing reconstruction for the 1-d signal can contribute to the communication of autonomous driving, intelligent robots, and fire exploration robots. To address the issue that fully connected layers in the LISTA method lack the ability to extract non-local features, this paper primarily designs a non-local fully connection layer and...
Show More
-
Review Article
Data Science and Machine Learning for Cyber Intrusion Detection: A Systematic Review
Issue:
Volume 11, Issue 1, June 2026
Pages:
8-21
Received:
28 February 2026
Accepted:
11 March 2026
Published:
18 March 2026
Abstract: The escalating sophistication and volume of cyberattacks have driven an urgent demand for intelligent Intrusion Detection Systems (IDS) that leverage Data Science (DS) and Machine Learning (ML). Despite rapid advances, existing reviews often focus narrowly on specific aspects without integrating the full data science and machine learning lifecycle. This paper presents a systematic review of DS and ML applications in cyber intrusion detection, covering 153 studies published from 2009 to 2025. The review systematically surveys benchmark datasets, data preprocessing and feature engineering techniques, classical ML and Deep Learning (DL) models, ensemble and hybrid strategies, class imbalance handling, and evaluation methodologies. A unified four-axis taxonomy is proposed to classify the literature, including learning strategy, imbalance handling, explainability level, and deployment context. A quantitative meta-analysis reveals that UNSW-NB15 and CIC-IDS2017 dominate at 71% combined dataset usage, deep learning represents 40% of algorithmic approaches, and only 34% of studies report per-class recall for minority attack types. Nine technically grounded research gaps are identified, spanning preprocessing standardization, cross-dataset evaluation, minority-class recall optimization, adversarial robustness, online and edge deployment, explainability for Security Operations Center (SOC) operations, federated learning, transformer and Large Language Model (LLMs) application, and zero-shot adaptation. The review further identifies eight emerging trends including attention-based and transformer architectures, LLMs, Graph Neural Networks (GNNs), federated and privacy-preserving learning, adversarial robustness, Explainable AI (XAI), zero-shot and few-shot detection, and Internet of Things (IoT) edge-based IDS. A seven-stage actionable architecture is proposed that integrates adaptive preprocessing, contrastive feature learning, recall-aware ensemble detection, XAI decision support, continual learning, and federated aggregation. This review provides researchers and practitioners with a structured roadmap for advancing the next generation of intelligent cyber intrusion detection systems.
Abstract: The escalating sophistication and volume of cyberattacks have driven an urgent demand for intelligent Intrusion Detection Systems (IDS) that leverage Data Science (DS) and Machine Learning (ML). Despite rapid advances, existing reviews often focus narrowly on specific aspects without integrating the full data science and machine learning lifecycle....
Show More
-
Research Article
Machine Learning Applications in Construction Supply Chain Management for Effective Project Delivery
Arinloye Grace Oshioname,
John Wasiu,
Ibrahim Abdulrazaq Olayinka*
Issue:
Volume 11, Issue 1, June 2026
Pages:
22-36
Received:
3 February 2026
Accepted:
19 March 2026
Published:
13 April 2026
Abstract: The construction industry faces significant challenges such as poor supplier communication and delayed deliveries in supply chain management (SCM), leading to project delays and cost overruns. This study investigates the application of machine learning (ML) to enhance the effectiveness of construction supply chain management for improved project delivery in Nigeria. A comprehensive methodology was employed, beginning with a literature review to identify key SCM factors, followed by a structured survey of 150 construction professionals to gather data on practices and project outcomes. The collected data was analyzed using the Classification Learner app in MATLAB, where various algorithms, including Decision Trees, Support Vector Machines (SVM), and ensemble methods, were trained and validated. Results indicated that Decision Trees (30%) and SVM (26.7%) were the most utilized and effective models for analyzing SCM data. The trained ML model achieved prediction accuracies of up to 90.7% in categorizing factors affecting project delivery. Key influential factors identified include supplier integration, inventory management, and logistics coordination. The study concludes that ML classification techniques are powerful tools for diagnosing SCM inefficiencies and predicting project performance. The findings provide a data-driven framework for construction stakeholders to prioritize SCM strategies, thereby mitigating risks and fostering more effective project delivery.
Abstract: The construction industry faces significant challenges such as poor supplier communication and delayed deliveries in supply chain management (SCM), leading to project delays and cost overruns. This study investigates the application of machine learning (ML) to enhance the effectiveness of construction supply chain management for improved project de...
Show More
-
Research Article
Gradient Boosting Revisited: Comparative Analysis of Selected Advances on Real-World Tabular Data
Issue:
Volume 11, Issue 1, June 2026
Pages:
37-52
Received:
26 April 2026
Accepted:
9 May 2026
Published:
12 June 2026
Abstract: Gradient Boosting has become one of the approaches design to improve general predictive performance as well as overcome some specific learning challenges. Though mature, there are still new adaptive variants being created to enhance flexibility, efficiency, as well as overall predictive power. However, there are limited benchmarking studies that sought to establish the generalisation abilities of these techniques especially the newer variants under varying conditions. This study, therefore, conducts a systematic analysis of seven Gradient Boosting models: XGBoost, LightGBM, CatBoost, HistGradientBoosting, GradientBoosting, AdaBoost, and the adaptive MorphBoost on ten benchmark datasets different challenges. All models were trained using a fixed 80:20 train–test split, with 3-fold cross-validation performed solely on the training portion to estimate stability. Performance was measured using accuracy, F1-score, and ROC-AUC to guarantee fairness and reproducibility. The findings indicate that CatBoost produced the highest mean accuracy of 0.9400 and a near-perfect ROC-AUC of 0.9915, which means that it can effectively generalize across diverse data types. HistGradientBoosting is identified as the most stable model across datasets with a good level of performance and computational efficiency, and it is currently followed by LightGBM and XGBoost. MorphBoost shows promise on binary and high-dimensional datasets where its implementation is fully supported, though its current lack of native multiclass handling limits general applicability. Generally, the research confirms that there is no single model that fits all circumstances; rather, dataset characteristics directly influence model performance. These results offer real-world guidance on the choice of boosting models and point to the areas where future research, particularly in adaptive and hybrid boosting techniques can be used to further enhance performance and generalization.
Abstract: Gradient Boosting has become one of the approaches design to improve general predictive performance as well as overcome some specific learning challenges. Though mature, there are still new adaptive variants being created to enhance flexibility, efficiency, as well as overall predictive power. However, there are limited benchmarking studies that sough...
Show More
-
Research Article
Predicting Girls’ Orientation Toward Technical Tracks in the Republic of Guinea Using Machine Learning:
A Comparative Study of Models and Determinants Analysis
Issue:
Volume 11, Issue 1, June 2026
Pages:
53-62
Received:
24 May 2026
Accepted:
8 June 2026
Published:
26 June 2026
DOI:
10.11648/j.mlr.20261101.15
Downloads:
Views:
Abstract: The purpose of this study is to develop a machine learning model capable of predicting girls’ orientation toward technical tracks in the Republic of Guinea. The dataset was constructed from the university placement records of high school graduates and includes academic and socio-demographic variables related to students’ orientation toward technical and non-technical tracks. In addition, the study identifies the variables associated with these orientations. To achieve this goal, four supervised learning algorithms, including Logistic Regression, Decision Trees, Random Forest, and Support Vector Machines (SVM) were used. The evaluation of the algorithms’ performance was based on the metrics Accuracy, Precision, F1-score, Recall, and AUC. The results show that the Random Forest model performs best, with an accuracy of 87.7%, an F1-score of 82.1%, and an AUC ROC of 0.954. Analysis of the variables reveals that the overall average score is the primary factor guiding girls toward technical tracks. This research highlights the importance of machine learning methods as a decision-making tool for policies aimed at education and the promotion of technical tracks among girls. Models were evaluated using an independent test set and a 5-fold stratified cross-validation procedure to assess robustness.
Abstract: The purpose of this study is to develop a machine learning model capable of predicting girls’ orientation toward technical tracks in the Republic of Guinea. The dataset was constructed from the university placement records of high school graduates and includes academic and socio-demographic variables related to students’ orientation toward technica...
Show More
