The early detection and classification of skin cancer are pivotal in improving patient outcomes and reducing healthcare burdens. However, traditional deep learning models in dermatological diagnostics often struggle with the nuanced differentiation of skin lesions. This paper introduces an approach to integrate an Advanced Heat Flow Layer into deep learning architectures for skin cancer classification, this method is centered on the principles of anisotropic diffusion, distinguishing itself from conventional image processing techniques by selectively smoothing image areas while preserving critical edge details, essential for accurate lesion identification. In our research, we utilized the Ham10000 dataset, enriched with data augmentation to simulate real-world variability, we conducted a comprehensive comparison of our model, featuring the Advanced Heat Flow Layer, against several benchmark deep learning models, including Sobel Edge Detection Layer. Our model, integrated with various layers of DenseNet121, consistently outperformed these benchmarks across key metrics such as accuracy, precision, recall, F1 score, and AUC, particularly with augmented data, this indicates a significant enhancement in the model’s ability to generalize and maintain critical diagnostic features under diverse conditions. Our code is available at, https://github.com/sanadv/SkinCancerClassificationModels/blob/main/Models.ipynb

The rapid expansion of medical data, characterized by its complex high-dimensional attributes, presents numerous promising opportunities and substantial challenges in healthcare analytics. Adopting effective feature selection techniques is essential to take advantage of the potential of such data. This research presents a modified algorithm called (mDA), which is the hybrid algorithm between the Evolutionary Population Dynamics and the Dragonfly Algorithm. This method combines Evolutionary Population Dynamics’s strength with the Dragonfly Algorithm’s flexible capabilities, offering a robust evolutionary machine learning approach specifically designed for medical data analysis. By integrating the dynamic population modeling of Evolutionary Population Dynamics with the adaptive search techniques of Dragonfly Algorithm, the proposed mDA significantly improves accuracy, reduces the number of features, and obtains the minimum average of the fitness scores. Comparative experiments conducted on seven diverse medical datasets against other established algorithms confirm the superior performance of the proposed mDA, establishing it as a valuable approach in examining complex medical data.

Education is a fundamental sector in all countries, where in some countries students com-pete to get an educational grant due to its high cost. The incorporation of artificial intelli-gence in education holds great promise for the advancement of educational systems and pro-cesses. Educational data mining involves the analysis of data generated within educational environments to extract valuable insights into student performance and other factors that enhance teaching and learning. This paper aims to analyze the factors influencing students' performance and consequently, assist granting organizations in selecting suitable students in the Arab region (Jordan as a use case). The problem was addressed using a rule-based tech-nique to facilitate the utilization and implementation of a decision support system. To this end, three classical rule induction algorithms, namely PART, JRip, and RIDOR, were em-ployed. The data utilized in this study was collected from undergraduate students at the University of Jordan from 2010 to 2020. The constructed models were evaluated based on metrics such as accuracy, recall, precision, and f1-score. The findings indicate that the JRip algorithm outperformed PART and RIDOR in most of the datasets based on f1-score metric. The interpreted decision rules of the best models reveal that both features; the average study years and high school averages play vital roles in deciding which students should receive scholarships. The paper concludes with several suggested implications to support and en-hance the decision-making process of granting agencies in the realm of higher education.

Autism spectrum disorder (ASD) is a developmental disorder that encompasses difficulties in communication (both verbal and non-verbal), social skills, and repetitive behaviors. The diagnosis of autism spectrum disorder typically involves specialized procedures and techniques, which can be time-consuming and expensive. The accuracy and efficiency of the diagnosis depend on the expertise of the specialists and the diagnostic methods employed. To address the growing need for early, rapid, cost-effective, and accurate diagnosis of autism spectrum disorder, there has been a search for advanced smart methods that can automatically classify the disorder. Machine learning offers sophisticated techniques for building automated classifiers that can be utilized by users and clinicians to enhance accuracy and efficiency in diagnosis. Eye-tracking scan paths have emerged as a tool increasingly used in autism spectrum disorder clinics. This methodology examines attentional processes by quantitatively measuring eye movements. Its precision, ease of use, and cost-effectiveness make it a promising platform for developing biomarkers for use in clinical trials for autism spectrum disorder. The detection of autism spectrum disorder can be achieved by observing the atypical visual attention patterns of children with the disorder compared to typically developing children. This study proposes a deep learning model, known as T-CNN-Autism Spectrum Disorder (T-CNN-ASD), that utilizes eye-tracking scans to classify participants into ASD and typical development (TD) groups. The proposed model consists of two hidden layers with 300 and 150 neurons, respectively, and underwent 10 rounds of cross-validation with a dropout rate of 20%. In the testing phase, the model achieved an accuracy of 95.59%, surpassing the accuracy of other machine learning algorithms such as random forest (RF), decision tree (DT), K-Nearest Neighbors (KNN), and multi-layer perceptron (MLP). Furthermore, the proposed model demonstrated superior performance when compared to the findings reported in previous studies. The results demonstrate that the proposed model can accurately classify children with ASD from those with TD without human intervention.

Predicting student performance is a critical aspect of educational systems. Although forecasting a student’s future performance is essential in many applications, it is a challenging process due to various factors. Previous research in this area has mainly focused on comparing machine learning methods to automate student evaluation and predict their final performance. However, there have been limited studies that thoroughly explore the issue of class imbalance using a deep learning approach. Moreover, the large dataset targeting university students makes it well-suited for in-depth analysis and increases the likelihood of obtaining more accurate results. This study presents a deep learning model based on convolution and introduces a comprehensive exploration of oversampling and undersampling methods to address the issue of imbalanced classes. The paper investigates various features and characteristics of undergraduate students at the University of Jordan, utilizing a large dataset collected from the university’s registration unit. These features include demographic information, attributes related to students’ majors, faculties, registrations, courses taken (such as passed, repeated, and completed), as well as their high school averages and performance in the first four semesters. The results demonstrate that the model performs exceptionally well in terms of gmean when predicting students’ excellence. This research project has significant implications and provides valuable insights to the research community and higher education managers, aiding in the development of improved strategies to enhance educational performance. Future researchers can utilize the methods employed in this paper during the data preprocessing stages and implement the demonstrated balancing strategies for further advancements in this field of study.

The communication revolution has perpetually reshaped the means through which people send and receive information. Social media is an important pillar of this revolution and has brought profound changes to various aspects of our lives. However, the open environment and popularity of these platforms inaugurate windows of opportunities for various cyber threats, thus social networks have become a fertile venue for spammers and other illegitimate users to execute their malicious activities. These activities include phishing hot and trendy topics and posting a wide range of contents in many topics. Hence, it is crucial to continuously introduce new techniques and approaches to detect and stop this category of users. This article proposes a novel and effective approach to detect social spammers. An investigation into several attributes to measure topic-dependent and topic-independent users’ behaviours on Twitter is carried out. The experiments of this study are undertaken on various machine learning classifiers. The performance of these classifiers is compared and their effectiveness is measured via a number of robust evaluation measures. Furthermore, the proposed approach is benchmarked against state-of-the-art social spam and anomalous detection techniques. These experiments report the effectiveness and utility of the proposed approach and embedded modules.

The Covid-19 vaccination process faced many problems; "who gets vaccinated" and "where is the nearest vaccination center" are two questions that entail a multi-Criteria problem. Such problems arise in the most inopportune time of the Covid-19 pandemic. The matching process depended on many variables: the current health state of the person, insurance company, location, and residency. Such variables are not static and differ from one country to another. Hence, this paper presents a new Multi-Criteria Hybrid Matching Method (MCHMM). MCHMM is a hybrid matching method extended from three multi-Criteria decision-making methods. MCHMM uses the strengths of three methods: Analytic Hierarchy Process (AHP), and Gale-Shapley algorithm. To fully automate the decision-making process of matching the MCHMM, which relies on CRITIC to develop the weight of Criteria, the weights are then used in AHP to develop the preference list of the Gale-Shapley algorithm (GS). Then the GS algorithm will produce the matched sets, eliminating the human factor and interference like favoritism in deciding. Such elimination is essential to reduce long waiting queues, proclivities, tendencies, biases, preconceptions, prejudices, and predispositions. Hence, reason and logic will guide the decision process, ridding it of impurities. MCHMM will match the person with the vaccination center most suited to her/his case, not based on distance only but also based on the Criteria s/he chooses. MCHMM is a unique hybrid algorithm that solves a multi-Criteria problem taking advantage of other algorithms. Furthermore, MCHMM can be generalized and can be used in any similar situation. Keywords : Covid-19, Gale-Shapley Algorithm, Analytic Hierarchy Process, Criteria Importance through Inter Criteria Correlation, Multi-Criteria Hybrid Matching Method DOI: https://doi.org/10.35741/issn.0258-2724.58.3.29

Source code authorship attribution is the task of identifying who develops the code based on learning based on the programmer style. It is one of the critical activities which used extensively in different aspects such as computer security, computer law, and plagiarism. This paper attempts to investigate source code authorship attribution by capturing natural language aspects of the code rather than only using minimal set of syntactic and stylistic code features as explored in the previous literature. It proposes an evolutionary feature selection model to improve the accuracy of authorship attribution by implementing two language models (uni-gram and bi-gram). The proposed approach uses K-Nearest Neighbor as a classifier and Genetic Algorithm as a feature selection technique. Two experiments have been demonstrated on a public Authorship Attribution dataset on GitHub, the experiments include various evolutionary feature selection models. Notably, the obtained results in both experiments were compared with the related studies, and show a significant improvement in terms of accuracy.

Type 2 diabetes is a common life-changing disease that has been growing rapidly in recent years. According to the World Health Organization, approximately 90% of patients with diabetes worldwide have type 2 diabetes. Although there is no permanent cure for type 2 diabetes, this disease needs to be detected at an early stage to provide prognostic support to allied health professionals and develop an effective prevention plan. This can be accomplished by analyzing medical datasets using data mining and machine-learning techniques. Due to their efficiency, metaheuristic algorithms are now utilized in medical datasets for detecting chronic diseases, with better results than traditional methods. The main goal is to improve the performance of the existing approaches for the detection of type 2 diabetes. A bio-inspired metaheuristic algorithm called cuttlefish was used to select the essential features in the medical data preprocessing stage. The performance of the proposed approach was compared to that of a well-known bio-inspired metaheuristic feature selection algorithm called the genetic algorithm. The features selected from the cuttlefish and genetic algorithms were used with different classifiers. The implementation was applied to two datasets: the Pima Indian diabetes dataset and the hospital Frankfurt diabetes dataset; generally, these datasets are asymmetry, but some of the features in these datasets are close to symmetry. The results show that the cuttlefish algorithm has better accuracy rates, particularly when the number of instances in the dataset increases.

The surge in cyber-attacks has driven demand for robust Intrusion detection systems (IDSs) to protect underlying data and sustain availability of network services. Detecting and classifying multiple type of attacks requires robust machine learning approaches that can analyze network traffic and take appropriate measures. Traffic data usually consists of redundant, irrelevant, and noisy information, which could have a negative influence on the model performance. In this paper, we propose an improved bio-inspired meta-heuristic algorithm for efficient detection and classification of multi-stage attacks. The proposed model uses a one-versus-all sub-model based technique to deal with the multi-class classification problem. Each sub-model employs an enhanced Harris Hawk optimization with extreme learning machine (ELM) as the base classifier. This hierarchy produces the best subset of features per attack, along with optimized ELM's weights, which can improve the detection rate significantly. The proposed technique was tested against various meta-heuristic algorithms and multi-class classifiers using the UNSWNB-15 dataset. In seven different types of attacks, experimental results outperformed other existing methods in terms of decreasing the crossover-error rate and obtaining the best values for the G-mean measure.

The growth of data creates more analysis and mining challenges related to speed and accuracy. Feature selection (FS) is an optimization problem used as a preprocessing phase to reduce the data dimensionality while obtaining the best classification accuracy. FS removes redundant and irrelevant features and preserves the best informative features. Various meta-heuristic optimization algorithms were employed in the literature to solve the FS problem. This paper proposes an improved Harris hawk optimization algorithm called (IHHO) to find the optimal feature set for classification purposes in a wrapper-based environment. Three main improvements are obtained in the binary version of HHO. The first improvement is to speed up the convergence, which is implemented using the most informative features in population initialization. Both filter-based and wrapper-based techniques are used during the initialization phase. The second one is to ensure the global and local search and avoid trapping into local optima using the X-shaped transfer function. While the third one is using the extreme learning machine as the base classifier to guide the searching process, speed up the convergence, and improve the accuracy of the FS process. The proposed model was evaluated using 18 well-known UCI benchmarks and compared with traditional HHO, particle swarm optimization, gray wolf optimizer, grasshopper optimization algorithm, and five standard filter-based techniques. The experiment results prove the superior performance of the IHHO compared to other algorithms and methods presented in the literature.

Smart cities have become an influential concept in urban development. Smart cities and their ap-plications aim to maintain a high quality of life by using smart technologies and enhancing eco-nomic productivity. Previous systematic literature reviews have considered the development of smart cities and highlighted their applications and services. However, no prior studies have com-prehensively investigated smart cities in relation to emergencies. To this end, the current paper aims to provide a research agenda reviewing the relevant literature that touches on smart tech-nologies during emergencies like the ongoing COVID-19 pandemic. Based on a systematic methodology centred on text mining analysis, our research identified the following three themes: (a) emergency response, which covers emergency management, traffic and unmanned aerial ve-hicles, waste disposal, and contact tracing; (b) motivation and outcome, which includes such sub -themes as smart urbanism, quality of life and the economy; and (c) technology and data, which covers social media, machine learning, Internet of Things, data-driven applications, and object detection. We comprehensively discuss each theme and offer suggestions for future research.

Training an Artificial Neural Network (ANN) algorithm is not trivial, which requires optimizing a set of weights and biases that increase dramatically with the increasing capacity of the neural network resulting in such hard optimization problems. Essentially, over recent decades, stochastic search algorithms have shown remarkable abilities for addressing hard optimization problems. On the other hand, pragmatically, abundant real-world problems suffer from the imbalance problem, where the distribution of data varies considerably among classes resulting in more training biases and variances which degrades the performance of the learning algorithm. This paper introduces three stochastic and metaheuristic algorithms for training the Multilayer Perceptron (MLP) neural network to solve the problem of imbalanced classifications. The utilized algorithms are the Grey Wolf Optimization (GWO), Particle Swarm Optimization (PSO), and the Salp Swarm Algorithm (SSA). The proposed GWO-MLP, PSO-MLP, and SSA-MLP are trained based on different objective functions; accuracy, f1-score, and g-mean. Whereas, it is evaluated based on 10 benchmark imbalanced datasets. The results show an advantage for f1-score, and g-mean fitness functions over the accuracy when the datasets are imbalanced.

This paper proposes an enhanced feature selection (FS) approach to improve the classification tasks, taking into account data dimensionality as a significant criterion of the dataset. High dimensionality may cause serious problems in classification that degrade the performance of the classifier. Among these problems: generating complex models (overfitting), increasing the learning time, and including redundant and irrelevant features in the learning model. FS is a data mining technique to minimize the number of dimensions (features) by getting rid of redundant and irrelevant features. Meanwhile, FS tries to maximize the classification performance. As FS is an optimization problem, meta-heuristic optimization algorithms can take place to achieve superior results in solving such problems. This paper proposes the Moth Flame Optimization (MFO) algorithm to tackle the FS problem. A new initialization method called opposition-based is proposed. Furthermore, a new update strategy is proposed to alleviate the local minima. The comparative results find that the proposed approach improves the MFO performance and outperforms other similar approaches.

Renewable energy sources are considered ubiquitous and drive the energy revolution. Energy producers suffer from inconsistent electricity generation. They often struggled with the unpredictability of the weather. Thus, making it challenging to balance supply and demand. Technologies like artificial intelligence (AI) and machine learning are effective ways to forecast, distribute, and manage renewable photovoltaic (PV) solar supplies. AI will make the energy forecasting system more connected, intelligent, reliable, and sustainable. AI can innovate how energy is used and help find solutions for decarbonizing energy systems. There are potential advantages to total energy forecasting. AI can support the growth and integration of PV solar energy. The article’s main objective is to use AI to forecast the output consumed power of the Yarmouk University PV solar system in Jordan. The total actual yield is 5548.96 MW h, and the performance ratio (PR) is 95.73%. Many techniques are used to predict the consumed solar power. The random forest model obtains the best results of root mean squared error and mean absolute error are 172.07 and 68.7, respectively. This accurate prediction allows for the maximum use of solar power and the minimal use of grid power. This work guides the operators to learn trends embedded in Yarmouk University’s historical data. These understood trends can be used to predict the consumption of solar power output. Thus, the control system and grid operators have advanced knowledge of the expected consumption of solar power at each hour of the day.

Software engineering is one of the most significant areas, which extensively used in educational and industrial fields. Software engineering education plays an essential role in keeping students up to date with software technologies, products, and processes that are commonly applied in the software industry. The software development project is one of the most important parts of the software engineering course, because it covers the practical side of the course. This type of project helps strengthening students’ skills to collaborate in a team spirit to work on software projects. Software project involves the composition of software product and process parts. Software product part represents software deliverables at each phase of Software Development Life Cycle (SDLC) while software process part captures team activities and behaviors during SDLC. The low-expectation teams face challenges during different stages of software project. Consequently, predicting performance of such teams is one of the most important tasks for learning process in software engineering education. The early prediction of performance for low-expectation teams would help instructors to address difficulties and challenges related to such teams at earliest possible phases of software project to avoid project failure. Several studies attempted to early predict the performance for low-expectation teams at different phases of SDLC. This study introduces swarm intelligence -based model which essentially aims to improve the prediction performance for low-expectation teams at earliest possible phases of SDLC by implementing Particle Swarm Optimization-K Nearest Neighbours (PSO-KNN), and it attempts to reduce the number of selected software product and process features to reach higher accuracy with identifying less than 40 relevant features. Experiments were conducted on the Software Engineering Team Assessment and Prediction (SETAP) project dataset. The proposed model was compared with the related studies and the state-of-the-art Machine Learning (ML) classifiers: Sequential Minimal Optimization (SMO), Simple Linear Regression (SLR), Naïve Bayes (NB), Multilayer Perceptron (MLP), standard KNN, and J48. The proposed model provides superior results compared to the traditional ML classifiers and state-of-the-art studies in the investigated phases of software product and process development.

Stress is a pensive issue in our competitive world and it has a huge impact on physical and mental health. Severe health issues may arise due to long exposure of stress. Hence, its timed detection can be helpful in managing stress periods. In this regard, electroencephalogram (EEG) based techniques have been widely explored, as stress severely impact the functioning and structure of brain. These non-invasive methods for stress detection need improvement in terms of predictive accuracy and reliability. In this work, a novel approach for stress detection has been presented using short duration of EEG signal. Entropy based features were extracted from EEG signal decomposed using stationary wavelet transform. Selected features were used for classification using different supervised machine learning algorithms. Further, different evolutionary inspired approaches were deployed to optimize the parameters of support vector machines (SVM) and perform feature weighting, simultaneously. SVM optimized using whale optimization algorithm resulted in an accuracy of 97.2559%. Accurate detection using short duration EEG signal shows potential of this technique for timed and reliable detection of stress.

Cyber-attacks and unauthorized application usage have increased due to the extensive use of Internet services and applications over computer networks, posing a threat to the service’s availability and consumers’ privacy. A network Intrusion Detection System (IDS) aims to detect aberrant traffic behavior that firewalls cannot detect. In IDSs, dimension reduction using the feature selection strategy has been shown to be more efficient. By reducing the data dimension and eliminating irrelevant and noisy data, several bio-inspired algorithms have been employed to improve the performance of an IDS. This paper discusses a modified bio-inspired algorithm, which is the Grey Wolf Optimization algorithm (GWO), that enhances the efficacy of the IDS in detecting both normal and anomalous traffic in the network. The main improvements cover the smart initialization phase that combines the filter and wrapper approaches to ensure that the informative features will be included in early iterations. In addition, we adopted a high-speed classification method, the Extreme Learning Machine (ELM), and used the modified GWO to tune the ELM’s parameters. The proposed technique was tested against various meta-heuristic algorithms using the UNSWNB-15 dataset. Because the generic attack is the most common attack type in the dataset, the primary goal of this paper was to detect generic attacks in network traffic. The proposed model outperformed other methods in minimizing the crossover error rate and false positive rate to less than 30%. Furthermore, it obtained the best results with 81%, 78%, and 84% for the accuracy, F1-score, and G-mean measures, respectively.

This survey is an effort to provide a research repository and a useful reference for researchers to guide them when planning to develop new Nature-inspired Algorithms tailored to solve Feature Selection problems (NIAs-FS). We identified and performed a thorough literature review in three main streams of research lines: Feature selection problem, optimization algorithms, particularly, meta-heuristic algorithms, and modifications applied to NIAs to tackle the FS problem. We provide a detailed overview of 156 different articles about NIAs modifications for tackling FS. We support our discussions by analytical views, visualized statistics, applied examples, open-source software systems, and discuss open issues related to FS and NIAs. Finally, the survey summarizes the main foundations of NIAs-FS with approximately 34 different operators investigated. The most popular operator is chaotic maps. Hybridization is the most widely used modification technique. There are three types of hybridization: Integrating NIA with another NIA, integrating NIA with a classifier, and integrating NIA with a classifier. The most widely used hybridization is the one that integrates a classifier with the NIA. Microarray and medical applications are the dominated applications where most of the NIA-FS are modified and used. Despite the popularity of the NIAs-FS, there are still many areas that need further investigation.