Explore the vast range of titles available.

Dataset size is considered a major concern in the medical domain, where lack of data is a common occurrence. This study aims to investigate the impact of dataset size on the overall performance of supervised classification models. We examined the performance of six widely-used models in the medical field, including support vector machine (SVM), neural networks (NN), C4.5 decision tree (DT), random forest (RF), adaboost (AB), and naïve Bayes (NB) on eighteen small medical UCI datasets. We further implemented three dataset size reduction scenarios on two large datasets and analyze the performance of the models when trained on each resulting dataset with respect to accuracy, precision, recall, f-score, specificity, and area under the ROC curve (AUC). Our results indicated that the overall performance of classifiers depend on how much a dataset represents the original distribution rather than its size. Moreover, we found that the most robust model for limited medical data is AB and NB, followed by SVM, and then RF and NN, while the least robust model is DT. Furthermore, an interesting observation is that a robust machine learning model to limited dataset does not necessary imply that it provides the best performance compared to other models.

Neonatal jaundice is a common condition worldwide. Failure of timely diagnosis and treatment can lead to death or brain injury. Current diagnostic approaches include a painful and time-consuming invasive blood test and non-invasive tests using costly transcutaneous bilirubinometers. Since periodic monitoring is crucial, multiple efforts have been made to develop non-invasive diagnostic tools using a smartphone camera. However, existing works rely either on skin or eye images using statistical or traditional machine learning methods. In this paper, we adopt a deep transfer learning approach based on eye, skin, and fused images. We also trained well-known traditional machine learning models, including multi-layer perceptron (MLP), support vector machine (SVM), decision tree (DT), and random forest (RF), and compared their performance with that of the transfer learning model. We collected our dataset using a smartphone camera. Moreover, unlike most of the existing contributions, we report accuracy, precision, recall, f-score, and area under the curve (AUC) for all the experiments and analyzed their significance statistically. Our results indicate that the transfer learning model performed the best with skin images, while traditional models achieved the best performance with eyes and fused features. Further, we found that the transfer learning model with skin features performed comparably to the MLP model with eye features.

Face gender recognition has many useful applications in human–robot interactions as it can improve the overall user experience. Support vector machines (SVM) and convolutional neural networks (CNNs) have been used successfully in this domain. Researchers have shown an increased interest in comparing and combining different feature extraction paradigms, including deep-learned features, hand-crafted features, and the fusion of both features. Related research in face gender recognition has been mostly restricted to limited comparisons of the deep-learned and fused features with the CNN model or only deep-learned features with the CNN and SVM models. In this work, we perform a comprehensive comparative study to analyze the classification performance of two widely used learning models (i.e., CNN and SVM), when they are combined with seven features that include hand-crafted, deep-learned, and fused features. The experiments were performed using two challenging unconstrained datasets, namely, Adience and Labeled Faces in the Wild. Further, we used T-tests to assess the statistical significance of the differences in performances with respect to the accuracy, f-score, and area under the curve. Our results proved that SVMs showed best performance with fused features, whereas CNN showed the best performance with deep-learned features. CNN outperformed SVM significantly at p < 0.05.

Financial planning plays an important role in people’s lives. The recent COVID-19 outbreak has caused sudden unemployment for many people across the globe, leaving them with a financial crisis. Recent surveys indicate that financial matters continue as the leading cause of stress for employees. Further, many millennials overspend and make unfortunate financial decisions due to their incapability to manage their earnings, which forbids them from maintaining financial satisfaction. Financial well-being as defined by The American Consumer Financial Protection Bureau (CFPB) is a state where one fully meets current and ongoing financial obligations, feels secure in their financial future, and is able to make choices to enjoy life. This work proposes a Personal Finance Management (PFM) system with a new architecture that aims to guide users to reach the state of financial well-being, as defined by CFPB. The proposed system consists of a rule-based system that provides users with actionable advice to make informed spending decisions and achieve their financial goals.

The novel coronavirus Severe Acute Respiratory Syndrome (SARS)-Coronavirus-2 (CoV-2) has resulted in an ongoing pandemic and has affected over 200 countries around the world. Mathematical epidemic models can be used to predict the course of an epidemic and develop methods for controlling it. As social contact is a key factor in disease spreading, modeling epidemics on contact networks has been increasingly used. In this work, we propose a simulation model for the spread of Coronavirus Disease 2019 (COVID-19) in Saudi Arabia using a network-based epidemic model. We generated a contact network that captures realistic social behaviors and dynamics of individuals in Saudi Arabia. The proposed model was used to evaluate the effectiveness of the control measures employed by the Saudi government, to predict the future dynamics of the disease in Saudi Arabia according to different scenarios, and to investigate multiple vaccination strategies. Our results suggest that Saudi Arabia would have faced a nationwide peak of the outbreak on 21 April 2020 with a total of approximately 26 million infections had it not imposed strict control measures. The results also indicate that social distancing plays a crucial role in determining the future local dynamics of the epidemic. Our results also show that the closure of schools and mosques had the maximum impact on delaying the epidemic peak and slowing down the infection rate. If a vaccine does not become available and no social distancing is practiced from 10 June 2020, our predictions suggest that the epidemic will end in Saudi Arabia at the beginning of November with over 13 million infected individuals, and it may take only 15 days to end the epidemic after 70% of the population receive a vaccine.

Online communication platforms face security and privacy challenges, especially in broad ecosystems, such as online social networks, where users are unfamiliar with each other. Consequently, employing trust management systems is crucial to ensuring the trustworthiness of participants, and thus, the content they share in the network. WhatsApp is one of the most popular message-based online social networks with over one billion users worldwide. Therefore, it is considered an attractive platform for cybercriminals who spread malware to gain unauthorized access to users’ accounts to steal their data or corrupt the system. None of the few trust management systems proposed in the online social network literature have considered WhatsApp as a use case. To this end, this paper introduces WhatsTrust, a trust management system for WhatsApp that evaluates the trustworthiness of users. A trust value accompanies each message to help the receiver make an informed decision regarding how to deal with the message. WhatsTrust is extensively evaluated through a strictly controlled empirical evaluation framework with two well-established trust management systems, namely EigenTrust and trust network analysis with subjective logic (TNA-SL) algorithms, as benchmarks. The experimental results demonstrate WhatsTrust’s dominance with respect to the success rate and execution time.

Peer-to-peer (P2P) architecture is increasingly gaining attention as a potential solution for the scalability problem facing the Internet of Things (IoT). It can be adopted for the fog computing layer to sustain the massive flow of data from constrained IoT nodes to the cloud. The success of a P2P-based system is entirely dependent on the continuity of resource sharing among individual nodes. Free riding is a severe problem that contradicts this main principle of P2P systems. It is understood that peers tend to consume resources from other peers without offering any in return. This free riding behavior can decrease system scalability and content availability, resulting in a decline in performance. Significant efforts have been made to hinder this behavior and to encourage cooperation amongst peers. To this end, we propose AFMIA, an Adjusted Free-Market-Inspired Approach that considers resources as goods that have dynamic prices based on the amount of supply and demand. Peers have wealth that can be increased by providing resources and spent by consuming them. The experimental results indicate that the proposed algorithm can successfully improve fairness without compromising on success rates.

Pilgrimage is one of the largest mass gatherings, where millions of Muslims gather annually from all over the world to perform Hajj. The stoning ritual during Hajj has been historically vulnerable to serious disasters that often cause severe impacts ranging from injuries to death tolls. In efforts to minimize the number and extent of the disasters, the stoning area has been expanded recently. However, no research has been carried out to study the evacuation effectiveness of the current exit placements in the area, which lies at the heart of effective minimization of the number and extent of the disasters. Therefore, this paper presents an in-depth study on emergency evacuation planning for the extended stoning area. It presents a simulation model of the expanded stoning area with the current exit placement. In addition, we suggested and examined four different exit placements considering evacuation scenarios in case of no hazard as well as two realistic hazard scenarios covering fire and bomb hazards. The simulation studied three stoning phases, beginning of stoning, during the peak hour of stoning, and ending of stoning at three scales of population sizes. The performance was measured in the light of evacuation time, percentage of evacuees, and percentage of crowd at each exit. The experimental results revealed that the current exits are not optimally positioned, and evacuation can be significantly improved through introducing a few more exits, or even through changing positions of the current ones.

Multi-agent systems are a common paradigm for building distributed systems in different domains such as networking, health care, swarm sensing, robotics, and transportation. Systems are usually designed or adjusted in order to reflect the performance trade-offs made according to the characteristics of the mission requirement. Research has acknowledged the crucial role that communication plays in solving many performance problems. Conversely, research efforts that address communication decisions are usually designed and evaluated with respect to a single predetermined performance goal. This work introduces Goal-Driven Communication, where communication in a multi-agent system is determined according to flexible performance goals. This work proposes an evolutionary approach that, given a performance goal, produces a communication strategy that can improve a multi-agent system's performance with respect to the desired goal. The evolved strategy determines what, when, and to whom the agents communicate. The proposed approach further enables tuning the trade-off between the performance goal and communication cost, to produce a strategy that achieves a good balance between the two objectives, according the system designer's needs.