Search

Recently, artificial intelligence (AI) has been one of the hottest topics in the technological world. Although it is involved in many domains, it was recently involved in the healthcare sector. AI can be used for diagnostics, drug development, treatment personalization, gene editing, disease prediction, and many more. It helps to improve healthcare services by benefiting medical professionals, hospitals, and patients. Saudi Arabia has a particular interest in the healthcare sector, and it has a clear vision for the future, which points toward the development of AI-based technologies. Few studies investigated the use of AI in Saudi healthcare, and most of them focused on healthcare employees' perceptions. This study is beyond the focus of the existing works. It aims at: 1) presenting the main AI-based healthcare applications; 2) exploring the use of AI in the Saudi healthcare sector; 3) addressing their ethical and legal challenges, along with the policy questions in Saudi healthcare; 4) studying the benefits of these AI-based applications and the acceptance of professionals to use AI in daily practice; 5) introducing the new Personal Data Protection Law (PDPL) in Saudi Arabia; and 6) discussing the importance of AI to the future of Saudi healthcare. To this purpose, a survey was distributed among four main Saudi hospitals. The findings showed that AI should not only lead to better health but also save manpower and simplify the healthcare processes. The respondents agreed that AI helps reflect human intellectual competencies and pushes its limits. Doi: 10.28991/ESJ-2022-06-04-05 Full Text: PDF

Exam proctoring is a hectic task i.e., the monitoring of students’ activities becomes difficult for supervisors in the examination rooms. It is a costly approach that requires much labor. Also, it is a difficult task for supervisors to keep an eye on all students at a time. Automatic exam activities recognition is therefore necessitating and a demanding field of research. In this research work, categorization of students’ activities during the exam is performed using a deep learning approach. A new deep CNN architecture with 46 layers is proposed which contains the characteristics of deep AlexNet and SqueezeNet. The model is engineered first with slight modifications in AlexNet. After then, the squeezed branch-like structure of SqueezeNet is grafted/embedded at two locations in the modified AlexNet architecture. The model is named as L2-GraftNet because of dual grafting blocks. The proposed model is first converted to a pre-trained model by performing its training with SoftMax classifier on the CIFAR-100 dataset. Afterwards, the features of the dataset prepared for exam activities categorization are extracted from the above-mentioned pre-trained model. The extracted features are then fed to the atom search optimization (ASO) approach for features optimization. The optimized features are passed to different variants of SVM and KNN classifiers. The best performance results attained are on the Fine KNN classifier with an accuracy of 93.88%. The satisfactory results prove the robustness of the proposed framework. Also, the proposed categorization provides a base for automated exam proctoring without the need for proctors in the exam halls.

Arabic documents are massively rising due to numerous contents utilized in websites, social media, and news articles. The classification of such documents in labelled categories is a significant and vital task that deserves more attention. Arabic Text Classification is an emerging research theme in Arabic Natural Language Processing. Recently, Deep Neural Network approaches have successfully been applied to many text classification problems, especially in English Text Classification. Convolutional Neural Network (CNN) is one of the best popular models. However, CNN is not highly applied in Arabic Text Classification. In addition, the recent studies did not achieve a high classification accuracy due to parameter setting issue. To overcome this limitation, a new hybrid classification model for Arabic Text is developed. This paper proposes Genetic Algorithms based Convolutional Neural Network for Arabic Text Classification. Genetic Algorithm is used to optimize the CNN parameters. The proposed model is tested using two large datasets and compared with the state-of-the art studies. The results showed that the classification accuracy achieved an improvement of 4 to 5%.

Automated recitation plays an important role in improving self-learning. It is based on Speech/Text recognition. The research in Arabic speech recognition is very limited. The few existing applications are only based on the Holy Qur’an. This article proposed a new system (Samee’a - ) to facilitate memorizing any kind of text such that poems, speeches and the Holy Qur’an. Samee’a system is based on Google Cloud Speech Recognition API to convert the Arabic speech to text and Jaro Winkler Distance algorithm to determine the similarity between the original and converted texts. The system has been tested using 70 collected files ranging between 12 to 400 words and some chapters from the Holy Qur’an. The average similarity achieved 83.33% for the 70 files and 69% for the selected chapters of the Holy Qur’an. These results were enhanced to 91.33 % and 95.66% after applying preprocessing operations on the text files and the Holly Qur’an respectively. To validate the obtained results, two comparison studies were performed. The Jaro Winker distance was successfully compared to the cosine and the Euclidean distance. In addition, the proposed system outperformed the related work with an improvement of the similarity reaching 5% when using section 30 of the Holy Qur’an. Finally, the user experience testing was carried out by 10 users of different ages (between 5 and 50-year-old) using small texts and some small chapters of the Holy Qur’an. The proposed system proved its efficiency.

Many countries are suffering from the COVID19 pandemic. The number of confirmed cases, recovered, and deaths are of concern to the countries having a high number of infected patients. Forecasting these parameters is a crucial way to control the spread of the disease and struggle with the pandemic. This study aimed at forecasting the number of cases and deaths in KSA using time-series and well-known statistical forecasting techniques including Exponential Smoothing and Linear Regression. The study is extended to forecast the number of cases in the main countries such that the US, Spain, and Brazil (having a large number of contamination) to validate the proposed models (Drift, SES, Holt, and ETS). The forecast results were validated using four evaluation measures. The results showed that the proposed ETS (resp. Drift) model is efficient to forecast the number of cases (resp. deaths). The comparison study, using the number of cases in KSA, showed that ETS (with RMSE reaching 18.44) outperforms the state-of-the art studies (with RMSE equal to 107.54). The proposed forecasting model can be used as a benchmark to tackle this pandemic in any country.

At the end of 2019, humans have been contracted by a new coronavirus, known as Covid-19, that caused a pneumonia outbreak in the Wuhan city in China. As this virus expanded globally outside China, the World Health Organization declared the outbreak as a global pandemic in March 2020. The paper aims to investigate the severity of the outbreak in different regions of Saudi Arabia and the impact of the lockdown through data visualization of different epidemiological metrics of Covid-19, such as the spread growth rate and the case fatality rate. The Data Visualization of the pandemic spread, and fatality was provided with Tableau using the real-time data.

Breast Cancer is a common disease among females. Early detection of the Breast Cancer aids in an easier efficient treatment. The application of Machine Learning algorithms can help in the diagnosis of this disease. There are three main problems related to Breast Cancer. The existing works focused only on one problem. In addition, the resulted accuracy still needs improvement. This research paper aims to identify the Breast Cancer diagnosis, predict the recurrence of the disease, and predict the survivability of its patients. This is achieved by using the Feedforward Neural Network (FFN) on the SEER (Surveillance, Epidemiology, and End Results) dataset by using different attributes and preprocessing of data for each problem. The obtained FFN classification accuracy resulted in 99.8% for the Breast Cancer diagnosis, 88.1% for the Breast Cancer recurrence, and 97.3% for the survivability.

Air travel demand is a crucial part of planning for airlines and airports. It helps in elaborating decisions and recognizing risks and opportunities. Forecasting air passenger demand is an interesting research study that deserves investigation. This problem requires prediction techniques such that Linear Regression and Neural Network. These techniques are efficient, but they have several parameters that necessitate appropriate values to provide the least error rate of prediction. Some recent air travel demand studies investigated Genetic Algorithms to provide optimal values for these parameters. In this article, we propose to explore the Firefly Algorithm (FA) and Particle Swarm Optimization (PSO) to find the optimal values for Linear Regression (LR) coefficients. This study presents two new hybrid prediction techniques (PSO based LR and FA based LR) to handle airline demand forecasting, which researchers have not previously covered. The results of PSO based LR, FA-based LR and LR are compared to find the best model with the lowest prediction error rate. The results showed that PSO based LR achieved the best prediction results with a lower error rate compared to FA based LR and LR alone. This study is performed using the data of Los Angeles International airport.

Most biometric identification applications suffer from the curse of dimensionality as the database size becomes very large, which could negatively affect both the identification performance and speed. In this paper, we use Projection Pursuit (PP) methods to determine clusters of individuals. Support Vector Machine (SVM) classifiers are then applied on each cluster of users separately. PP clustering is conducted using Friedman and Kurtosis projection indices optimized by Genetic Algorithm and Particle Swarm Optimization methods. Experimental results obtained using YALE face database showed improvement in the performance and speed of face identification system

Arabic natural language processing (ANLP) consists of developing techniques and tools that can utilize and analyze the Arabic language in both written and spoken contexts. ANLP makes an important contribution to many existing developed systems. It provides Arabic and non-Arabic speakers with helpful and convenient tools that can be used in different domains. Modern ANLP tools are developed using machine learning (ML) techniques. ML algorithms are widely used in NLP because of their high accuracy rate regardless of the robustness of the data that is used and because of the ease with which they can be implemented. On the other hand, the methodology of ANLP applications based on ML involves several distinct phases. It is, therefore, crucial to recognize and understand these phases in detail as well as the most widely used ML algorithms. This survey discusses this concept in detail, shows the involvement of ML techniques in developing such tools, and identifies well-known techniques used in ANLP. Moreover, this survey discusses the characteristics and complexity of the Arabic language in addition to the importance and needs of ANLP.

Natural language processing (NLP) captured the attention of researchers for the last years. NLP is applied in various applications and several disciplines. Arabic is a language that also benefited from NLP. However, only few Arabic datasets are available for researchers. For that, applying the Arabic NLP is limited in these datasets. Hence, this paper introduces a new dataset, SNAD. SNAD is collected to fill the gap in Arabic datasets, especially for classification using deep learning. The dataset has more than 45,000 records. Each record consists of the news title, news details, in addition to the news class. The dataset has six different classes. Moreover, cleaning and preprocessing are applied to the raw data to make it more efficient for classification purpose. Finally, the dataset is validated using the Convolutional Neural Networks and the result is efficient. The dataset is freely available online.

With the development of technology, e-commerce became an essential part of an individual’s life, where individuals could easily purchase and sell products over the internet. However, fraud attempts; specifically credit card fraudulent attacks are rapidly increasing. Cards may potentially be stolen; fake records are being used and credit cards are subject to being hacked. Artificial Intelligence techniques tackle these credit card fraud attacks, by identifying patterns that predict false transactions. Both Machine Learning and Deep Learning models are used to detect and prevent fraud attacks. Machine Learning techniques provide positive results only when the dataset is small and do not have complex patterns. In contrast, Deep Learning deals with huge and complex datasets. However, most of the existing studies on Deep Learning have used private datasets, and therefore, did not provide a broad comparative study. This paper aims to improve the detection of credit card fraud attacks using Long Short-Term Memory Recurrent Neural Network (LSTM RNN) with a public dataset. Our proposed model proved to be effective. It achieved an accuracy rate of 99.4% which is higher compared to other existing Machine and Deep Learning techniques.

This article consists of using biologically inspired algorithms in order to detect potentially interesting structures in large and multidimensional data sets. Data exploration and the detection of interesting structures are based on the use of Projection Pursuit that involves the definition and the optimization of an index associated with each direction or projection. The optimization of a projection index should provide a set of multiple optima that is expected to correspond to interesting graphical representations in low dimensional space. The implementation of the bio-inspired algorithms along with the projection pursuit develops a new software called EPP-Lab. Projection pursuit is widely used in different scientific domains (biology, pharmacy, bioinformatics, biometry, etc) but not widely present in the well-known softwares. EPP-Lab is dedicated to recognize and visualize clusters and outlying observations on one dimension from high dimensional and multivariate data sets. It includes different statistical techniques for results analysis. It provides several features and gives the user the option to adjust the parameters of the selected bio-inspired methods or to use defaults values. EPP-Lab is a unique software for detection, visualization and analysis of non-linear structures. The performance of this tool has been validated by testing different real and simulated data sets.

The development of Recommender Systems in universities students’ portal provides a significant impact on the student study plan and grades. Such systems are more beneficial when they match the software interface usability goals. Henceforth, the students will be more likely to use them. To the best of our knowledge, there are no proposed prototypes neither research works for Courses Recommender Systems in universities students’ portal. Hence, this paper will present a prototype proposal for such a system. The proposed prototype evaluation achieved 4.8643 on the Likert scale reflecting a system with high usability.

Cancer is one of the leading causes of deaths in the last two decades. It is either diagnosed malignant or benign - depending upon the severity of the infection and the current stage. The conventional methods require a detailed physical inspection by an expert dermatologist, which is time-consuming and imprecise. Therefore, several computer vision methods are introduced lately, which are cost-effective and somewhat accurate. In this work, we propose a new automated approach for skin lesion detection and recognition using a deep convolutional neural network (DCNN). The proposed cascaded design incorporates three fundamental steps including; a) contrast enhancement through fast local Laplacian filtering (FlLpF) along HSV color transformation; b) lesion boundary extraction using color CNN approach by following XOR operation; c) in-depth features extraction by applying transfer learning using Inception V3 model prior to feature fusion using hamming distance (HD) approach. An entropy controlled feature selection method is also introduced for the selection of the most discriminant features. The proposed method is tested on PH2 and ISIC 2017 datasets, whereas the recognition phase is validated on PH2, ISBI 2016, and ISBI 2017 datasets. From the results, it is concluded that the proposed method outperforms several existing methods and attained accuracy 98.4% on PH2 dataset, 95.1% on ISBI dataset and 94.8% on ISBI 2017 dataset.

In the recent years, Arabic Natural Language Processing, including Text summarization, Text simplification, Text Categorization and other Natural Language-related disciplines, are attracting more researchers. Appropriate resources for Arabic Text Categorization are becoming a big necessity for the development of this research. The few existing corpora are not ready for use, they require preprocessing and filtering operations. In addition, most of them are not organized based on standard classification methods which makes unbalanced classes and thus reduced the classification accuracy. This paper proposes a New Arabic Dataset (NADA) for Text Categorization purpose. This corpus is composed of two existing corpora OSAC and DAA. The new corpus is preprocessed and filtered using the recent state of the art methods. It is also organized based on Dewey decimal classification scheme and Synthetic Minority Over-Sampling Technique. The experiment results show that NADA is an efficient dataset ready for use in Arabic Text Categorization.

The university Examination Timetabling Problem is one of the most important scheduling problems that numerous educational organizations have to find out a solution to manage their examinations by assigning these events to particular timeslots, rooms and/or invigilators. This problem is NP-complete due to the number of conflicting constraints that must be considered in the resolution. Particle Swarm Optimization (PSO) technique is a common intelligent method that has been successfully applied to many hard combinatorial optimization problems. The purpose of this paper is to expose a number of articles that appeared this last decade and used the PSO technique to solve the University examination timetabling problem. The overall techniques are described, focusing on the particle representation and updating. This research also offers insight into how well the PSO algorithm performs compared with other algorithms used to solve the same problem and datasets. Finally, a summary of the described algorithms and their most distinguishing features is presented in addition to future research directions.

Timetabling is a discrete and combinatorial optimization problem. It is usually handled with heuristic and/or artificial intelligent approaches. Examination timetabling problem is part of Timetabling problem that consists of allocating a set of exams (courses) in a finite period of time while satisfying a set of constraints of various types. Examination timetabling problem have still been studying due to their wide applications. Particle Swarm Optimization (PSO) technique is a well-known intelligent method that has been successfully applied to many hard combinatorial optimization problems but not widely used in this field of problem. In this paper, we propose a new hybrid approach based on PSO. The proposed algorithm has considerable results and usually a good stability.

BeamGA is a general hybrid heuristic framework that can be used to solve the median problem in comparative genomics, where any distance function can be used. It starts with a heuristic search approach (local beam search) in order to generate a number of solutions. Then a Genetic Algorithm (GA) is applied to refine the solutions. It considers true biological evolution scenarios by applying the concept of common intervals during the GA optimization process.

In modern cloud data centers, tens of thousands of virtual machines are monitored and managed every day. Cloud service providers task is revealed to be complex and time consuming. In order to make it easier, researchers focused on minimizing the number of virtual machine instances to be monitored in fine grain. This is achieved by clustering virtual machines in classes of similar behavior. Projection Pursuit is one well-known clustering technique that yields good results in this domain. This paper presents a first study of clustering virtual machines using Projection Pursuit technique. The technique is tested on simulated instances, the experiments show promising results.

For most real-world biometric identification applications, the training database size could be very large, i.e. in the range of several thousands. This yields to the curse of dimensionality problem. The downside of such a problem is that it could negatively affect both the identification performance and speed. In this paper we use Exploratory Projection Pursuit (EPP) methods to determine clusters of users having significant similarities and then apply Support Vector Machine (SVM) classifiers on each cluster of users independently. This allows reducing the dimensionality of the dataset for training SVMs and thus improving the performance of user identification.

Data mining is a huge and an interesting domain of patterns extraction. Spatial clustering has been distinguished as a major data mining task. It consists of grouping comparable objects into classes while taking into account the spatial aspect. It plays an important role in different areas, it is why several techniques have been proposed. This article presents a new approach based on the search of spatial clusters using Projection Pursuit with a dual mode. The idea is to look for projections revealing clusters that take into account the spatial information contained in the data. This involves solving a bi-objective problem where the first objective function is a projection index dedicated to the search of clusters and the second objective is a distance function defined for this purpose. Accordingly, a Multiobjective bio-inspired algorithm is used. Combining the spatial aspect with the Projection Pursuit and introducing a Multiobjective bio-inspired method in the same context is a first study in the literature. This new approach has been tested with real and simulated datasets, the experiments yield promising results.

Detecting outliers in the context of multivariate data is known as an important but difficult task and there already exist several detection methods. Most of the proposed methods are based either on the Mahalanobis distance of the observations to the center of the distribution or on a projection pursuit (PP) approach. In the present paper we focus on the one-dimensional PP approach which may be of particular interest when the data are not elliptically symmetric. We give a survey of the statistical literature on PP for multivariate outliers etection and investigate the pros and cons of the different methods. We also propose the use of a recent heuristic optimization algorithm called Tribes for multivariate outliers detection in the projection pursuit context.

In order to summarize and represent graphically multidimensional data in statistics, projection pursuit methods look for projection axes which reveal structures, such as possible groups or outliers, by optimizing a function called projection index. To determine these possible interesting structures, it is necessary to choose an optimization method capable to find not only the global optimum of the projection index but also the local optima susceptible to reveal these structures. For this purpose, we suggest a metaheuristic which does not ask for many parameters to settle and which provokes premature convergence to local optima. This method called Tribes is a hybrid Particle Swarm Optimization method (PSO) based on a stochastic optimization technique developed in [2]. The computation is fast even for big volumes of data so that the use of the method in the field of projection pursuit fulfills the statistician expectations.