Ilyas, Mohammad

Epilepsy is a multifaceted neurological disorder characterized by superfluous and recurrent seizure activity. Electroencephalogram (EEG) signals are indispensable tools for epilepsy diagnosis that reflect real-time insights of brain activity. Recently, epilepsy researchers have increasingly utilized Deep Learning (DL) architectures for early and timely diagnosis. This research focuses on resolving the challenges, such as data diversity, scarcity, limited labels, and privacy, by proposing potential contributions for epilepsy detection, prediction, and forecasting tasks without impacting the accuracy of the outcome. The proposed design of diversity-enhanced data augmentation initially averts data scarcity and inter-patient variability constraints for multiclass epilepsy detection. The potential features are extracted using a graph theory-based approach by analyzing the inherently dynamic characteristics of augmented EEG data. It utilizes a novel temporal weight fluctuation method to recognize the drastic temporal fluctuations and data patterns realized in EEG signals. Designing the Siamese neural network-based few-shot learning strategy offers a robust framework for multiclass epilepsy detection.

Human Activity Recognition (HAR) plays a crucial role in various applications, including healthcare, fitness tracking, security, and smart environments, by enabling the automatic classification of human actions based on sensor and visual data. This dissertation presents a comprehensive exploration of HAR utilizing machine learning, sensor-based data, and Fusion approaches. HAR involves classifying human activities over time by analyzing data from sensors such as accelerometers and gyroscopes. Recent advancements in computational technology and sensor availability have driven significant progress in this field, enabling the integration of these sensors into smartphones and other devices. The first study outlines the foundational aspects of HAR and reviews existing literature, highlighting the importance of machine learning applications in healthcare, athletics, and personal use. In the second study, the focus shifts to addressing challenges in handling large-scale, variable, and noisy sensor data for HAR systems. The research applies machine learning algorithms to the KU-HAR dataset, revealing that the LightGBM classifier outperforms others in key performance metrics such as accuracy, precision, recall, and F1 score. This study underscores the continued relevance of optimizing machine learning techniques for improved HAR systems. The study highlights the potential for future research to explore more advanced fusion techniques to fully leverage different data modalities for HAR. The third study focuses on overcoming common challenges in HAR research, such as varying smartphone models and sensor configurations, by employing data fusion techniques.

The Internet of Things (IoT) has undergone remarkable expansion in recent years, leading to a proliferation of devices capable of connecting to the internet, collecting data, and sharing information. However, this rapid growth has also introduced a myriad of security challenges, resulting in an uptick in cyber-attacks targeting IoT infrastructures. To mitigate these threats and ensure the integrity of data, researchers have been actively engaged in the development of robust Intrusion Detection Systems (IDS) utilizing various machine learning (ML) techniques. This dissertation presents a comprehensive overview of three distinct approaches toward IoT intrusion detection, each leveraging ML methodologies to enhance security measures. The first approach focuses on a multi-class classification algorithm, integrating models such as random forest, logistic regression (LR), decision tree (DT), and Xgboost. Through meticulous evaluation utilizing evaluation metrics including F1 score, recall, and precision under the Receiver Operating Characteristics (ROC) curve, this approach demonstrates a remarkable 99 % accuracy in detecting IoT attacks. In the second approach, a deep ensemble model comprising Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM) architectures is proposed for intrusion detection in IoT environments. Evaluation on the UNSW 2018 IoT Botnet dataset showcases the proficiency of this approach, achieving an accuracy of 98.4 % in identifying malicious activities. Lastly, the dissertation explores a real-time Intrusion Detection System (IDS) framework deployed within the Pyspark architecture, aimed at efficiently detecting IoT attacks while minimizing detection time.

The Internet of Things (IoT) refers to a network of interconnected nodes constantly engaged in communication, data exchange, and the utilization of various network protocols. Previous research has demonstrated that IoT devices are highly susceptible to cyber-attacks, posing a significant threat to data security. This vulnerability is primarily attributed to their susceptibility to exploitation and their resource constraints. To counter these threats, Intrusion Detection Systems (IDS) are employed. This study aims to contribute to the field by enhancing IDS detection efficiency through the integration of Ensemble Learning (EL) methods with traditional Machine Learning (ML) and deep learning (DL) models. To bolster IDS performance, we initially utilize a binary ML classification approach to classify IoT network traffic as either normal or abnormal, employing EL methods such as Stacking and Voting. Once this binary ML model exhibits high detection rates, we extend our approach by incorporating a ML multi-class framework to classify attack types. This further enhances IDS performance by implementing the same Ensemble Learning methods. Additionally, for further enhancement and evaluation of the intrusion detection system, we employ DL methods, leveraging deep learning techniques, ensemble feature selections, and ensemble methods. Our DL approach is designed to classify IoT network traffic. This comprehensive approach encompasses various supervised ML, and DL algorithms with ensemble methods. The proposed models are trained on TON-IoT network traffic datasets. The ensemble approaches are evaluated using a comprehensive metrics and compared for their effectiveness in addressing this classification tasks. The ensemble classifiers achieved higher accuracy rates compared to individual models, a result attributed to the diversity of learning mechanisms and strengths harnessed through ensemble learning. By combining these strategies, we successfully improved prediction accuracy while minimizing classification errors. The outcomes of these methodologies underscore their potential to significantly enhance the effectiveness of the Intrusion Detection System.

In recent years, many protocols for efficient Multicasting have been proposed.
However, many of the Internet Service Providers (ISPs) are reluctant to use multicastenabled
routers in their networks. To provide such incentives, new protocols are
needed to improve the quality of their services. The challenge is to find a compromise
between allocating Bandwidth (BW) among different flows in a fair manner, and
favoring multicast sessions over unicast sessions. In addition, the overall higher level
of receiver satisfaction should be achieved.
In this dissertation, we propose three new innovative protocols to favor
multicast sessions over unicast sessions. Multicast Favored BW Allocation-
Logarithmic (MFBA-Log) and Multicast Favored BW Allocation-Linear (MFBALin)
protocols allocate BW proportional to the number of down stream receivers.
The proposed Multicast Reserved BW Allocation (MRBA) protocol allocates part of the BW in the links only to multicast sessions. Simulation results show the increase in
the overall level of Receiver Satisfaction in the network.

The purpose of this research was to assess how prepared Florida's State University System (SUS) institutions have been during the past five years (2008-2013) in responding to the challenges of globalization. The research also established institutional trends for the past five years (2008-2013) and projections for the next five years to seize the opportunities offered by globalization and to produce graduates with global competency skills. Ten of the 12 SUS institutions studied in this research were Florida A&M University (FAMU), Florida Atlantic University (FAU), Florida Gulf Coast University (FGCU), Florida International University (FIU), Florida State University (FSU), University of Central Florida (UCF), University of Florida (UF), University of North Florida (UNF), University of South Florida (USF), and University of West Florida (UWF). The research was conducted as a case study using multi-method approach. The quantitative analysis was based on the information collected from the institutions and from the integrated postsecondary education data system (IPEDS). The qualitative analysis was based on the institutional mission statements, vision statements, and strategic plans. The quantitative analysis used six data parameters to compute a globalization composite index (GCI) for institutional comparisons and for establishing trends and future projections. Integrating quantitative and qualitative analyses led to the research findings of this study. Based on this study, the institutional preparedness for globalization has been low for six SUS institutions (FAMU, FAU, FGCU, UCF, UNF, and UWF) and has been medium for the remaining four (FIU, FSU, UF, and USF). The trend analysis showed that institutional preparedness could be improved significantly if robust and focused efforts are made over the next five years. In that case, the institutional preparedness for FAMU, FGCU, UNF, and UWF could ascend to medium; for FAU and UCF, it could improve to medium+; and for FIU, FSU, UF, and USF, it could reach high. The research concluded with some recommendations to help the leadership of Florida and the SUS institutions in responding effectively to the challenges of globalization. A few recommendations for future research in this field also are provided.

The growing demand for faster connection to the Internet service and wireless
multimedia applications has motivated the development of broadband wireless access
technologies in recent years. WiMAX has enabled convergence of mobile and fixed
broadband networks through a common wide-area radio-access technology and flexible
network architecture. Scheduling is a fundamental component in resource management in
WiMAX networks and plays the main role in meeting QoS requirements such as delay,
throughput and packet loss for different classes of service. In this dissertation work, the performance of uplink schedulers at the fixed WiMAX MAC layer has been considered, we proposed an Adaptive Hierarchical Weighted Fair Queuing Scheduling algorithm, the new scheduling algorithm adapts to changes in traffic, at the same time; it is able to heuristically enhance the performance of WiMAX network under most circumstances. The heuristic nature of this scheduling algorithm enables the MAC layer to meet the QoS requirements of the users. The performance of this adaptive WiMAX Uplink algorithm has been evaluated by simulation using MATLAB. Results indicate that the algorithm is efficient in scheduling the Base Stations’ traffic loads, and improves QoS. The utilization of relay stations is studied and simulation results are compared with the case without using relay stations. The results show that the proposed scheduling algorithm improves Quality of Service of WiMAX system.

The purpose of this thesis is to study three types of
bandwidth allocation strategies for a network integrating
voice and data, commonly referred to as Integrated Services
Digital Network or ISDN, using GPSS V simulati ons. The
strategies are Non-Boundary, Movable-Boundary, and Non- and
Movable- Boundary with Digital Speech Interpolation. The
theoretical behavior of each strategy is discussed. Exact
solutions for small systems with one or two slots is shown
along with approximations for larger systems. General
descriptions of the GPSS models for each strategy is
provided. The GPSS model source code for each strategy is
presented in the Appendix. Simulation is used to explore
the effects of the service time ratio a of voice and data on
system performance. Also, the performance of Time Division
Multiplex or TDM systems with 24 channels, commonly referred
to as Tl, is shown.

In the present computer age, cellular technology and portable computers are becoming an integral part of the life. Each computer user wants to access the computing resources, irrespective of the location. Because of this need the computing paradigm "Mobile Computing" has assumed a primary role in modern computer communication technology. While dimensioning the network resources, it is very important to know how the users move around the geographical area covered by the cellular network. This knowledge allows us to plan the system resources in order to achieve the QoS required. The major factors that affect the performance, along with the mobility pattern of the mobile user, are the speed at which the user is moving and the load on the network. In this research, we study different types of mobility patterns the user can follow and it's impact on the network services. We have proposed and evaluated a reservation scheme to improve the QoS in the cellular network. The reservation scheme reserves some part of the bandwidth for handoff connections. We have developed simulation programs and have studied three mobility patterns namely leading movement type, random motion, and square-street mobility pattern for measuring the QoS for cellular network. It has been observed from the results that at an average speed of 50 miles per hour with the average loading of the network, a significant improvement in QoS has been achieved for all the mobility patterns by using the reservation scheme.

High speed ATM networks support a variety of communication services, that have different traffic characteristics, which causes the network to be congested quickly. An ATM network with different communication services, data, voice and video, is simulated to study the effect of congestion on network operation. A modified leaky bucket mechanism is used to shape the traffic entering the network, which improved the performance in terms of cell losses and cell delay. The original leaky bucket mechanism is so conservative, that it drops a large number of ATM cells. Another scheme called virtual leaky bucket is proposed in this thesis. In this scheme violating cells are marked and then allowed to enter the network. The scheme is simulated and its performance is compared to the leaky bucket mechanism. Shaped virtual leaky bucket with marking is shown to have much better performance as long as the minimum requirements of non-violating cells are guaranteed.