This study proposes a novel machine learning (ML) approach for early detection of thyroid disorders using data mining and ensemble learning techniques. By leveraging a diverse dataset—including demographic details, medical history, symptoms, and diagnostic test results—a high‐precision model is developed to predict the risk of thyroid dysfunction. The methodology integrates self‐adaptive stacking, weighted metafeatures, and Bayesian optimization to enhance model performance. After preprocessing, feature selection, and correlation analysis, various ensemble classifiers are trained and evaluated. Ensemble methods improve prediction accuracy by combining multiple base models, making them well suited for complex medical classification tasks. The proposed model achieved outstanding performance, with an accuracy of 99.46%, sensitivity (recall for class 1) of 99.85%, specificity (recall for class 0) of 94.83%, and an overall F1‐score of 99.71%. The slight variation in class‐wise F1‐scores reflects the impact of class imbalance, where metrics tend to favor the majority class. Nevertheless, these metrics collectively underscore the model’s robustness in effectively identifying at‐risk individuals while minimizing false positives and false negatives. Heat maps and other visualization tools were used to interpret patterns and improve model transparency. The findings support the potential of data‐driven approaches to enhance early diagnosis, inform personalized treatment plans, and improve clinical decision‐making. This work not only contributes to better thyroid disease detection but also advances the development of robust, interpretable, and scalable ML models for healthcare applications.

After the increased reliance on online education, online assessment became an essential tool for educators to remotely monitor and evaluate students’ understanding in order to assist them properly. However, the laborious process of creating exam questions is a challenge for most teachers. Thus, automated Question Generation aims to assist teachers by generating questions from given data. Limited research has been conducted to tackle this issue in the Arabic Language due to the complexity of the language and the limited amount of available Arabic data. This article explores different implementations of the transformer models, that demonstrated their superiority in natural language processing. Three approaches were introduced to tackle this problem with Arabic data using an Arabic-based transformer, an English-based transformer, and a multilingual-based transformer. Each of the fine-tuned models was trained using ARCD, XGLUE, DialectBench, and ArabicQA data sets and evaluated on automatic and manual metrics. Two of the proposed models achieve state-of-the-art results on the Arabic question generation task. The English transformer obtained a ROUGE score of 0.59 on XGLUE, while the Arabic transformer model achieves 0.49 on ARCD. Both of these models demonstrate excellent quality of questions through human-conducted evaluations by achieving low WER and high GC, U, and A scores.

Abstract POSIT is a numerical system consists of sign, regime, exponential, and fraction bits to overcome some limitations of the IEEE‐754 floating point (FP) representation. This work proposes a technique to hide critical information in the least significant bits (LSBs) of the POSIT FP representation by exploiting approximate computing (AC). The proposed technique, called information hiding using POSIT and LSB (IHUPL), explores the opportunities offered by both the POSIT representation and the AC to hide information with a minimum loss in accuracy and other performance metrics. IHUPL offers two options for hiding information: either by embedding one digit of each character into each pixel or by embedding all the digits of the character at the same time into each pixel of the alpha‐red, green, blue, or black/white image. Experimental results are evaluated for benchmark images and showed that IHUPL enhanced the accuracy of embedding data into LSB of POSIT FP by an average of 5%, the image quality improvement rates of IHUPL are 19% and 16% for options 1 and 2, respectively. Besides the encoding method using IHUPL, the paper outlines an extraction decoding technique that saves the original replacement bits in a key‐image to recover the hidden security message.

Infection with a virus can lead to a range of illnesses in humans, including cancer. When viruses infect a host, they may disrupt normal host function and cause deadly diseases. Understanding complicated viral illnesses requires novel viral genome prediction. Since many of the sequences in assembled contigs from human samples are not identical to known genomes, many assembled contigs are labeled “unknown” by conventional alignments. In this study, sequences from 19 metagenomic investigations were used to create the model proposed here, and these sequences were examined and classified using BLAST. We implemented k-mer counting and the bag-of-words technique using CountVectorizer. As far as we are aware, this work represents the first framework that combines natural language processing (NLP) along with traditional ML classification approaches on raw metagenomic contigs to automatically identify viruses in a variety of human biospecimens. The suggested models are general rather than specialized to a particular viral family. Since the proposed methodology is precise and simple, we may incorporate it into computer-aided diagnosis (CAD) systems to make day-to-day hospital activities easier. In the last stage, binary classification of deoxyribonucleic acid (DNA) with normal and viral genomes was performed using traditional ML classifiers. Using the KNN classifier, the suggested model achieved 98.6% classification accuracy along with 98.5% precision, 98.6% recall, 0.984 F1 score, 0.896 Mattews correlation coefficient, 0.895 kappa, 0.97 classification success index and detection rate of 98.6% for the prediction of viral genomes in DNA. Compared to previously developed ML techniques, the model achieved a significantly greater performance for viral genome prediction.

BACKGROUND: Here propose a computer-aided diagnosis (CAD) system to differentiate COVID-19 (the coronavirus disease of 2019) patients from normal cases, as well as to perform infection region segmentation along with infection severity estimation using computed tomography (CT) images. The developed system facilitates timely administration of appropriate treatment by identifying the disease stage without reliance on medical professionals. So far, this developed model gives the most accurate, fully automatic COVID-19 real-time CAD framework. RESULTS: The CT image dataset of COVID-19 and non-COVID-19 individuals were subjected to conventional ML stages to perform binary classification. In the feature extraction stage, SIFT, SURF, ORB image descriptors and bag of features technique were implemented for the appropriate differentiation of chest CT regions affected with COVID-19 from normal cases. This is the first work introducing this concept for COVID-19 diagnosis application. The preferred diverse database and selected features that are invariant to scale, rotation, distortion, noise etc. make this framework real-time applicable. Also, this fully automatic approach which is faster compared to existing models helps to incorporate it into CAD systems. The severity score was measured based on the infected regions along the lung field. Infected regions were segmented through a three-class semantic segmentation of the lung CT image. Using severity score, the disease stages were classified as mild if the lesion area covers less than 25% of the lung area; moderate if 25-50% and severe if greater than 50%. Our proposed model resulted in classification accuracy of 99.7% with a PNN classifier, along with area under the curve (AUC) of 0.9988, 99.6% sensitivity, 99.9% specificity and a misclassification rate of 0.0027. The developed infected region segmentation model gave 99.47% global accuracy, 94.04% mean accuracy, 0.8968 mean IoU (intersection over union), 0.9899 weighted IoU, and a mean Boundary F1 (BF) contour matching score of 0.9453, using Deepabv3+ with its weights initialized using ResNet-50. CONCLUSIONS: The developed CAD system model is able to perform fully automatic and accurate diagnosis of COVID-19 along with infected region extraction and disease stage identification. The ORB image descriptor with bag of features technique and PNN classifier achieved the superior classification performance.

Recent advances in machine learning have shown promising results for detecting network intrusion through supervised machine learning. However, such techniques are ineffective for new types of attacks. In the preferred unsupervised and semi-supervised cases, these newer techniques suffer from lower accuracy and higher rates of false alarms. This work proposes a machine learning model that combines auto-encoder with one-class support vectors machine. In this model, the auto-encoders learn the representation of the input data in a latent space and reduces the dimensionality of the input data. The dimensionality-reduced input is then extracted from the auto-encoder and passed to a one-class support vectors machine to classify the network event as an attack or a normal event. The model is trained on normal network events only. The proposed model is then evaluated and compared with several existing models. It achieves high accuracy when tested on the NSL-KDD and KDD99 datasets, with total accuracies of 96.24% and 99.45%, respectively.

Top-Rank-K Frequent Itemset (or Pattern) Mining (FPM) is an important data mining task, where user decides on the number of top frequency ranks of patterns (itemsets) they want to mine from a transactional dataset. This problem does not require the minimum support threshold parameter that is typically used in FPM problems. Rather, the algorithms solving the Top-Rank-K FPM problem are fed with K, the number of frequency ranks of itemsets required, to compute the threshold internally. This paper presents two declarative approaches to tackle the Top-Rank-K Closed FPM problem. The first approach is Boolean Satisfiability-based (SAT-based) where we propose an effective encoding for the problem along with an efficient algorithm employing this encoding. The second approach is CP-based, that is, utilizes Constraint Programming technique, where a simple CP model is exploited in an innovative manner to mine the Top-Rank-K Closed FPM itemsets from transactional datasets. Both approaches are evaluated experimentally against other declarative and imperative algorithms. The proposed SAT-based approach significantly outperforms IM, another SAT-based approach, and outperforms the proposed CP-approach for sparse and moderate datasets, whereas the latter excels on dense datasets. An extensive study has been conducted to assess the proposed approaches in terms of their feasibility, performance factors, and practicality of use.

There has recently been a significant growth in resource-constrained devices (RCDs) that exchange sensitive and private data. Lightweight ciphers are designed to implement confidentiality in RCDs. Hardware implementation of lightweight cipher should minimize resources, including area, power, and energy. Hardware Trojans (HTs) are malicious circuits that are inserted into designs, including lightweight ciphers, to modify cipher behavior and leak sensitive data. Although runtime monitoring is a very effective method to implement design-for-trust and detect hardware trojans, it requires significant resources and degrades performance. The main motivation of this research is to design RCD-friendly runtime monitoring and create a trusted design with minimal resource overhead. This article develops a low-power, low-energy and trusted design based on a smart runtime monitoring algorithm targeted for lightweight ciphers in RCDs, implemented in the FPGA platform. The algorithm is adaptive, minimizes resource, and maximizes confidence and HT detection. The novelties of the algorithm are its adaptive activation of checking and its validation. Checking is triggered when a predefined critical node is active, which limits power/energy overhead. Our proposed algorithm has adaptive positive aging because it forgoes validation when a critical node has been already proven safe. To optimize trust, the validation uses a reverse-function design. The implementation results show that proposed algorithm reduces area, power and energy, when compared with existing algorithms. The proposed algorithm achieves 37%–55% reduction in energy and power, and has an average of 53% improvement in LE×energy metric. Furthermore, our proposed algorithm reduces the area by 25% when both encryption and decryptions are implemented.

Abstract The early days of voice over IP (VoIP) adoption were characterized by a lack of concern and awareness about security issues related to its use. Indeed, service providers and users were mostly preoccupied with issues related to its quality, functionality, and cost. Now that VoIP is a mainstream communication technology, security has become a major issue. This paper investigates the major security threats for VoIP communications and proposes a multipath approach solution, especially targeted for low bandwidth networks. Results show that security has an effect on VoIP quality especially for a large distance between communicating nodes and packet size. Results also show that our proposed multipath solution reduces significantly packet losses and performs better than single routing techniques in networks with low bandwidth capacities. Copyright © 2017 John Wiley & Sons, Ltd.

Abstract In this paper, we propose a novel least significant bit embedding approach that capitalizes on the skewed distribution of letter and word frequencies to achieve higher image capacity, quality, and security. We initially conduct a study that involves all of the character frequencies using a data set of 14.245 billion characters. Huffman coding for each character is generated on the basis of its probability of occurrence. Furthermore, the top 100 000 most frequent words are transformed into a smaller ciphertext that has a lower cost. Our work demonstrates that recognizing characters and words on the basis of their frequency patterns and prioritizing them accordingly has a greater prospect of reducing the overall cost of embedding. The proposed scheme significantly outperforms Lempel–Ziv–Welch compression with an average of 45% fewer embedded bits. Moreover, the image quality is improved by a mean peak signal‐to‐noise ratio value of 6.9%. The proposed method also establishes a security embedding by proposing a novel shuffling algorithm. Copyright © 2017 John Wiley & Sons, Ltd.

Abstract In this paper we propose an automatic methodology to verify the soundness of model checking reduction techniques. The idea is to use the consistency of the specifications to verify if the reduced model is faithful to the original one. The user provides the reduction technique, the specification and the system under verification. Then, using Higher Order Logic he verifies automatically if the reduction technique is soundly applied. The method is completely defined in an MDG–HOL special integration platform that combines an automatic high level model checking tool Multiway Decision Graphs (MDGs) within the HOL theorem prover. We provide two case studies, the first one is the reduction using SAT–MDG of an Island Tunnel Controller and the second one is the MDG–HOL assume-guarantee reduction of the Look-Aside Interface. The obtained results of our approach offer a considerable gain in terms of the correctness of heuristics and reduction techniques as applied to commercial model checking, however a small penalty is paid in terms of CPU time and memory usage.

In this paper, we are interested in defining a platform for high level model checking using multiway decision graphs (MDGs) within high order logic. The platform is based on the logical formulation of an MDG as a directed formulae (DF). The DF is defined in the HOL theorem prover where the many sorted first-order logic is characterized as a HOL built-in data type. Then, the HOL inference rules are defined to check the well-formedness conditions of any directed formula. Based on this formalization, the MDGs operations are defined as inference rules and consistency and well-formedness proof of each operation is provided. Finally, some experimental results are presented to show the performance of the MDG-HOL platform. The obtained results show that this platform offers a considerable gain in terms of automation without sacrificing CPU time and memory usage.