Papers by Azween Abdullah
Computers, Jan 11, 2019
Criminal network activities, which are usually secret and stealthy, present certain difficulties ... more Criminal network activities, which are usually secret and stealthy, present certain difficulties in conducting criminal network analysis (CNA) because of the lack of complete datasets. The collection of criminal activities data in these networks tends to be incomplete and inconsistent, which is reflected structurally in the criminal network in the form of missing nodes (actors) and links (relationships). Criminal networks are commonly analyzed using social network analysis (SNA) models. Most machine learning techniques that rely on the metrics of SNA models in the development of hidden or missing link prediction models utilize supervised learning. However, supervised learning usually requires the availability of a large dataset to train the link prediction model in order to achieve an optimum performance level. Therefore, this research is conducted to explore the application of deep reinforcement learning (DRL) in developing a criminal network hidden links prediction model from the reconstruction of a corrupted criminal network dataset. The experiment conducted on the model indicates that the dataset generated by the DRL model through self-play or self-simulation can be used to train the link prediction model. The DRL link prediction model exhibits a better performance than a conventional supervised machine learning technique, such as the gradient boosting machine (GBM) trained with a relatively smaller domain dataset.
MATEC web of conferences, 2021
In the United States, the manufacturing ecosystem is rebuilt and developed through innovation wit... more In the United States, the manufacturing ecosystem is rebuilt and developed through innovation with the promotion of AMP 2.0. For this reason, the industry has spurred the development of 5G, Artificial Intelligence (AI), and Machine Learning (ML) technologies which is being applied on the smart factories to integrate production process management, product service and distribution, collaboration, and customized production requirements. These smart factories need to effectively solve security problems with a high detection rate for a smooth operation. However, number of security related cases occurring in the smart factories has been increasing due to botnet Distributed Denial of Service (DDoS) attacks that threaten the network security operated on the Internet of Things (IoT) platform. Against botnet attacks, security network of the smart factory must improve its defensive capability. Among many security solutions, botnet detection using honeypot has been shown to be effective in early studies. In order to solve the problem of closely monitoring and acquiring botnet attack behaviour, honeypot is a method to detect botnet attackers by intentionally creating resources within the network. As a result, the traced content is recorded in a log file. In addition, these log files are classified quickly with high accuracy with a support of machine learning operation. Hence, productivity is increase, while stability of the smart factory is reinforced. In this study, a botnet detection model was proposed by combining honeypot with machine learning, specifically designed for smart factories. The investigation was carried out in a hardware configuration virtually mimicking a smart factory environment.
International journal of multimedia and ubiquitous engineering, Sep 30, 2016
Proxy Mobile IPv6 (PMIPv6) was initially introduced to assist unicast network-based mobility. In ... more Proxy Mobile IPv6 (PMIPv6) was initially introduced to assist unicast network-based mobility. In recent years, new approaches have been introduced to provide multicast support in PMIPv6. IP multicast is an imperative mechanism for internet video provision. As the usage of internet data traffic remains to develop rapidly, there is a need to optimize and improve the performance of multicast service. Issues such as large overhead, high packet loss rate, single point of failure, service disruption time, handover latency, and non-route optimization needs to be tackled efficiently. To provide multicast services in PMIPv6, route optimization, global mobility, load balancing and context transfer approaches have been introduced. The foremost aim of this paper is to study and analyze these methods via qualitative analysis. This is to focus the advantages and the limitations of the current approaches.
Fabrication and Characterization Of Poly Lactic Acid (PLA)-Starch Based Bioplastic Composites
IOP conference series, Nov 1, 2019
In this study, to make a good bioplastic composite, starch-based bioplastic is produced by adding... more In this study, to make a good bioplastic composite, starch-based bioplastic is produced by adding polylactic acid (PLA) to improve its properties. PLA was added into starch-based bioplastic with various concentrations of 0, 3, and 10 wt.%. The extrusion was performed at 90-150 °C and compression moulding process was conducted at 150 °C and pressured at 50 kgf/cm2. Bioplastic composites have been characterized to know its properties. FTIR analysis indicated shifting and increasing spectra of interaction between PLA and starch-based bioplastic. Contact angle and solubility analysis revealed that adding PLA can increase the stability of hydrophobic characteristic and insoluble properties. The combination of PLA and starch-based bioplastic can improve the mechanical properties. In addition, thermal properties of bioplastic composites have a better thermal stability and produce a lower melting point thus the energy needed to melt for bioplastic composites becomes milt as raising PLA composition. The density of bioplastics was in the range of 1.2 - 1.3 g/cm3 that would be good for light bioplastic. The results of this study showed that the combination of starch-based bioplastics and PLA at low concentration (10wt.%) potentially could enhance the properties of bioplastic composites for food packaging.
World Academy of Science, Engineering and Technology, International Journal of Computer, Electrical, Automation, Control and Information Engineering, Mar 1, 2011
In this paper, cloud resource broker using goalbased request in medical application is proposed. ... more In this paper, cloud resource broker using goalbased request in medical application is proposed. To handle recent huge production of digital images and data in medical informatics application, the cloud resource broker could be used by medical practitioner for proper process in discovering and selecting correct information and application. This paper summarizes several reviewed articles to relate medical informatics application with current broker technology and presents a research work in applying goal-based request in cloud resource broker to optimize the use of resources in cloud environment. The objective of proposing a new kind of resource broker is to enhance the current resource scheduling, discovery, and selection procedures. We believed that it could help to maximize resources allocation in medical informatics application.
Intrusion Detection Systems are security programs to decide whether events and activities occurri... more Intrusion Detection Systems are security programs to decide whether events and activities occurring in a system or network are intrusive or legitimate. The objective of IDS is to identify intrusions with low false alarms and high detection rate while consuming lesser resources. There are plentiful issues in traditional IDS including regular updating, low detection capability to unknown attacks, non-adapting high false alarms rate, high resources consumption and many others. Similarly, Intelligent Network IDS have snags of performance efficiency, false positive and false negative while today's advance Neural Network approaches are also facing training/learning overhead, high false alarms and low detection rate. Soft computing is an innovative field to develop intelligent IDS while minimizing the deficiencies in other approaches. In this paper, an efficient soft computing approach is proposed by selecting an optimum subset of features. For training and testing of system, NSL-KDD dataset is preferred over KDD-Cup as there are approved deficits in KDD-Cup. Features transformation and optimum subset selection is done by Linear Discriminant Analysis (LDA) algorithm and Genetic Algorithm (GA) respectively. Radial Basis Function (RBF) is adapted as features classifier. Empirical results show that the new proposed system gives better and robust representation of an ideal intrusion detection system while having the reduced number of features, low false alarms, high detection rate and minimum computation cost.
Predictive modeling for intrusions in communication systems using GARMA and ARMA models
The strength of time series modeling is generally not used in almost all current intrusion detect... more The strength of time series modeling is generally not used in almost all current intrusion detection and prevention systems. By having time series models, system administrators will be able to better plan resource allocation and system readiness to defend against malicious activities. In this paper, we address the knowledge gap by investigating the possible inclusion of a statistical based time series modeling that can be seamlessly integrated into existing cyber defense system. Cyber-attack processes exhibit long range dependence and in order to investigate such properties a new class of Generalized Autoregressive Moving Average (GARMA) can be used. In this paper, GARMA (1,2; δ,1) model is fitted to cyber-attack data sets. Three different estimation methods are used to estimate the parameters. The Hannan-Rissanen Algorithm, Whittle Estimation Method and Maximum Likelihood Estimation methods are used to estimate the parameters of the GARMA (1,2;δ,1). Point forecasts to predict the attack rate possibly hours ahead of time also has been done and the performance of the models and estimation methods are discussed. The investigation of the case-study will confirm that by exploiting the statistical properties, it is possible to predict cyber-attacks (at least in terms of attack rate) with good accuracy. This kind of forecasting capability would provide sufficient early-warning time for defenders to adjust their defense configurations or resource allocations.
A Comparative Performance of Optimizers and Tuning of Neural Networks for Spoof Detection Framework
International Journal of Advanced Computer Science and Applications, 2022
Jurnal Sains Materi Indonesia, Apr 30, 2019
FABRICATION AND CHARACTERIZATION OF STARCH BASED BIOPLASTICS WITH PALM OIL ADDITION. In this work... more FABRICATION AND CHARACTERIZATION OF STARCH BASED BIOPLASTICS WITH PALM OIL ADDITION. In this work, starch-based bioplastics in advancing its properties were positively arranged with the addition of palm oil. Starch-based bioplastics were produced by dry blending method and compression technique with mixing starch and glycerol (3:1, w/w) then adding palm oil at various concentration (0%, 2.5%, 5% and 7.5% w/w). Morphology of bioplastics presented that palm oil wrapped bioplastics granules which influenced hydrophobicity properties of bioplastics compared by increasing contact angle of bioplastics from 45.95 o (0% of palm oil) to 61.98 o (5% of palm oil). This result indicated that the addition of palm oil could develop the properties of bioplastics to hold absorbing water molecules. Moreover, the melting point of bioplastics also affected shifting temperature from 115 o C to be 100 o C that could save the energy needed during heating process. FTIR analysis showed that C=O group at wavenumber 1747 cm-1 was dependable the interaction between starch-glycerol and palm oil. Furthermore, the addition of palm oil would accelerate the biodegradation process. Although the mechanical properties of bioplastics have not increased, the addition of palm oil on bioplastics fabrication is an alternative to improve the characteristic of bioplastics, especially physical, thermal, hydrophobicity and biodegradation properties.
IoT-enabled Industrial Wireless Sensor Networks for Next Generation Smart Grid Industry 4.0
ICSES transactions on computer networks and communications, Mar 30, 2018
The existing complex and aging electricity grid systems infrastructure due to lack of the user ut... more The existing complex and aging electricity grid systems infrastructure due to lack of the user utility interaction and one-way power flow is suffering from fraud detection, distribution automation, overload conditions, power quality issues, peak load management, energy loss, and power fault diagnostics. To meet the 21st-century energy requirements, next-generation power grid is envisioned to fully address these concerns in a sophisticated manner. Recently, the rise of new digital industrial technology Internet-of-Thing (IoT) and Industrial Wireless Sensor Networks (IWSNs)-based networking systems introduces the fourth stage of industrialization, commonly known as Industry 4.0. Recently, Industry 4.0, has lately gained a lot of interest from researchers, manufacturers, and application developers. Industry 4.0 will make it possible to produce higher-quality goods at reduced costs due to timely gathering and analyzing data across machines. This will increase manufacturing productivity and faster industrial growth in a more flexible manner which results in more economic benefits. This economics shift, in turn, will ultimately change the profile of people due to changing the competitiveness of companies and regions in the world. The wired or wireless integration of various components inside a factory to implement a flexible and reconfigurable manufacturing system, i.e., smart factory, is one of the key features of Industry 4.0.
International Journal of Advanced Computer Science and Applications, 2020
With the explosive growth of the Internet and the desire to harness the value of the information ... more With the explosive growth of the Internet and the desire to harness the value of the information it contains, the prediction of possible links (relationships) between key players in social networks based on graph-theory principles has garnered great attention in recent years. Consequently, many fields of scientific research have converged in the development of graph analysis techniques to examine the structure of social networks with a very large number of users. However, the relationship between persons within the social network may not be evident when the data-capture process is incomplete or a relationship may have not yet developed between participants who will establish some form of actual interaction in the future. As such, the link-prediction metrics for certain social networks such as criminal networks, which tend to have highly inaccurate data records, may need to incorporate additional circumstantial factors (metadata) to improve their predictive accuracy. One of the key difficulties in link-prediction methods is extracting the structural attributes necessary for the classification of links. In this research, we analysed a few key structural attributes of a network-oriented dataset based on proposed social network analysis (SNA) metrics for the development of link-prediction models. By combining structural features and metadata, the objective of this research was to develop a prediction model that leverages the deep reinforcement learning (DRL) classification technique to predict links/edges even on relatively small-scale datasets, which can constrain the ability to train supervised machine-learning models that have adequate predictive accuracy.
IEEE Access, 2020
Evidently, criminal network activities have shown an increasing trend in terms of complexity and ... more Evidently, criminal network activities have shown an increasing trend in terms of complexity and frequency, particularly with the advent of social media and modern telecommunication systems. In these circumstances, law enforcement agencies have to be armed with advance criminal network analysis (CNA) tools capable of uncovering with speed, probable key hidden relationships (links/edges) and players (nodes) in order to anticipate, undermine and cripple organised crime syndicates and activities. The development of link prediction models for network orientated domains is based on Social Network Analysis (SNA) methods and models. The key objective of this research is to develop a link prediction model that incorporates a fusion of metadata (i.e. environment data sources such as arrest warrants, judicial judgement, wiretap records and police station proximity) with a time-evolving criminal dataset in order to be aware of real-world situations to improve the quality of link prediction. Based on the review of related work, most of the models are constructed by leveraging on classical machine learning (ML) techniques such as support vector machine (SVM) without metadata fusion. The problem with the use of classical ML techniques is the lack of available domain dataset which is sufficiently large for training purpose. Compared to sociaI network, criminal network dataset by nature tends to relatively much smaller. In view of this, deep reinforcement learning (DRL) technique which could improve the training of models with the self-generated dataset is leveraged upon to construct the model. In this research, a purely time-evolving DRL model (TDRL-CNA) without metadata fusion is designed as a baseline for comparison with the metadata fusion model (FDRL-CNA). The experimental results show that the predictive accuracy of new and recurrent links by the FDRL-CNA model is higher than the baseline TDRL-CNA model that does not factor data fusion from different data sources. INDEX TERMS Data fusion, time-evolving network, criminal network analysis, deep reinforcement learning, node similarity.
Dust Emission Investigation of Agricultural Powdered Materials
Journal of physics, May 1, 2022
Towards self-organizing Service Oriented Architecture
Page 1. AbstractDevelopment of internet and World Wide Web technologies has enabled access to ma... more Page 1. AbstractDevelopment of internet and World Wide Web technologies has enabled access to many types of services over the web. Large and complex computational units can be built / composed out of the available services. ...
Load balancing techniques in cloud computing environment: A review
Journal of King Saud University - Computer and Information Sciences, Jul 1, 2022
Abstract Cloud Computing is a robust model that allows users and organizations to purchase requir... more Abstract Cloud Computing is a robust model that allows users and organizations to purchase required services per their needs. The model offers many services such as storage, platforms for deployment, convenient access to web services, and so on. Load Balancing is a common issue in the cloud that makes it hard to maintain the performance of the applications adjacent to the Quality of Service (QoS) measurement and following the Service Level Agreement (SLA) document as required from the cloud providers to enterprises. Cloud providers struggle to distribute equal workload among the servers. An efficient LB technique should optimize and ensure high user satisfaction by utilizing the resources of VMs efficiently. This paper presents a comprehensive review of various Load Balancing techniques in a static, dynamic, and nature-inspired cloud environment to address the Data Center Response Time and overall performance. An analytical review of the algorithms is provided, and a research gap is concluded for the future research perspective in this domain. This research also provides a graphical representation of reviewed algorithms to highlight the operational flow. Additionally, this review presents a fault-tolerant framework and explores the other existing frameworks in the recent literature.
Journal of King Saud University - Computer and Information Sciences, May 1, 2022
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Multicast Support in Network Based Mobility Management: Current Challenges and Solutions
Advanced Science Letters, Oct 1, 2016
Network based mobility management protocol can be termed as Proxy MIPv6 (PMIPv6). It is introduce... more Network based mobility management protocol can be termed as Proxy MIPv6 (PMIPv6). It is introduced as a unicast network based mobility management. Route optimization as well as context transfers approaches are some intra domain multicast support mechanisms in PMIPv6. As for inter domain multicast support mechanism on PMIPv6 domain, global mobility and load balancing are some of the approaches. Enabling multicast support mechanism in PMIPv6 is an important issue. This is because, it is an effective IP multicast communication. Moreover, nowadays application such as IPTV traffic densely populates IP multicast network concerning packet loss, service disruption time, and handover latency are at high priority. The objective of this paper is to quantitatively investigate and analyze intra and inter domain mechanisms for multicast support in PMIPv6. Using total number of signaling and handover latency as the analytical parameters, it also highlights the strengths, the limitations of these mechanisms. After analysis of the prevailing state of the art, an enhanced scheme has proposed in which both route optimization and context transfer techniques are integrated. The proposed scheme can brings entire benefits of the multicast service on PMIPv6 domain
Journal of King Saud University - Computer and Information Sciences, Dec 1, 2021
The scale of criminal networks (e.g. drug syndicates and terrorist networks) extends globally and... more The scale of criminal networks (e.g. drug syndicates and terrorist networks) extends globally and poses national security threat to many nations as they also tend to be technologically advance (e.g. Dark Web and Silk Road cryptocurrency). Therefore, it is critical for law enforcement agencies to be equipped with the latest tools in criminal network analysis (CNA) to obtain key hidden links (relationships) within criminal networks to preempt and disrupt criminal network structures and activities. Current hidden or missing link predictive models that are based on Social Network Analysis models rely on ML techniques to improve the performance of the models in terms of predictive accuracy and computing power. Given the improvement in the recent performance of Deep Reinforcement Learning (DRL) techniques which could train ML models through self-generated dataset, DRL can be usefully applied to domains with relatively smaller dataset such as criminal networks. The objective of this study is to assess the comparative performance of a CNA hidden link prediction model developed using DRL techniques against classical ML models such as gradient boosting machine (GBM), random forest (RF) and support vector machine (SVM). The experiment results exhibit an improvement in the performance of the DRL model of about 7.4% over the next best performing classical RF model trained within 1500 iterations. The performance of these link prediction models can be scaled up with the parallel processing capabilities of graphical processing units (GPUs), to significantly improve the speed of training the model and the prediction of hidden links.
Uploads
Papers by Azween Abdullah