Search

With growing demand on resources situated at the cloud datacenters, the need for customers’ resource selection techniques becomes paramount in dealing with the concerns of resource inefficiency. Techniques such as metaheuristics are promising than the heuristics, most especially when handling large scheduling request. However, addressing certain limitations attributed to the metaheuristic such as slow convergence speed and imbalance between its local and global search could enable it become even more promising for customers service selection. In this work, we propose a cloud customers service selection scheme called Dynamic Multi-Objective Orthogonal Taguchi-Cat (DMOOTC). In the proposed scheme, avoidance of local entrapment is achieved by not only increasing its convergence speed, but balancing between its local and global search through the incorporation of Taguchi orthogonal approach. To enable the scheme to meet customers’ expectations, Pareto dominant strategy is incorporated providing better options for customers in selecting their service preferences. The implementation of our proposed scheme with that of the benchmarked schemes is carried out on CloudSim simulator tool. With two scheduling scenarios under consideration, simulation results show for the first scenario, our proposed DMOOTC scheme provides better service choices with minimum total execution time and cost (with up to 42.87%, 35.47%, 25.49% and 38.62%, 35.32%, 25.56% reduction) and achieves 21.64%, 18.97% and 13.19% improvement for the second scenario in terms of execution time compared to that of the benchmarked schemes. Similarly, statistical results based on 95% confidence interval for the whole scheduling scheme also show that our proposed scheme can be much more reliable than the benchmarked scheme. This is an indication that the proposed DMOOTC can meet customers’ expectations while providing guaranteed performance of the whole cloud computing environment.

Wireless Sensors Networks have been the focus of significant attention from research and development due to their applications of collecting data from various fields such as smart cities, power grids, transportation systems, medical sectors, military, and rural areas. Accurate and reliable measurements for insightful data analysis and decision-making are the ultimate goals of sensor networks for critical domains. However, the raw data collected by WSNs usually are not reliable and inaccurate due to the imperfect nature of WSNs. Identifying misbehaviours or anomalies in the network is important for providing reliable and secure functioning of the network. However, due to resource constraints, a lightweight detection scheme is a major design challenge in sensor networks. This paper aims at designing and developing a lightweight anomaly detection scheme to improve efficiency in terms of reducing the computational complexity and communication and improving memory utilization overhead while maintaining high accuracy. To achieve this aim, one-class learning and dimension reduction concepts were used in the design. The One-Class Support Vector Machine (OCSVM) with hyper-ellipsoid variance was used for anomaly detection due to its advantage in classifying unlabelled and multivariate data. Various One-Class Support Vector Machine formulations have been investigated and Centred-Ellipsoid has been adopted in this study due to its effectiveness. Centred-Ellipsoid is the most effective kernel among studies formulations. To decrease the computational complexity and improve memory utilization, the dimensions of the data were reduced using the Candid Covariance-Free Incremental Principal Component Analysis (CCIPCA) algorithm. Extensive experiments were conducted to evaluate the proposed lightweight anomaly detection scheme. Results in terms of detection accuracy, memory utilization, computational complexity, and communication overhead show that the proposed scheme is effective and efficient compared few existing schemes evaluated. The proposed anomaly detection scheme achieved the accuracy higher than 98%, with O(nd) memory utilization and no communication overhead.

Intrusion detection systems (IDS) is an important security measure used to secure cloud resources, however, IDS often suffer from poor detection accuracy due to coordinated attacks such as a DDoS. Various research on distributed IDSs have been proposed to detect DDoS however, the limitations of these works the lack of technique to determine an appropriate period to share attack information among nodes in the distributed IDS. Therefore, this article proposes a distributed IDS that uses a binary segmentation change point detection algorithm to address the appropriate period to send attack information to nodes in distributed IDS and using parallel Stochastic Gradient Descent with Support Vector Machine (SGD-SVM) to achieve the distributed detection. The result of the proposed scheme was implemented in Apache Spark using NSL-KDD benchmark intrusion detection dataset. Experimental results show that the proposed distributed intrusion detection scheme outperforms existing distributed IDS for cloud computing.

Vehicular Ad Hoc Networks (VANETs) are among the main enablers for future Intelligent Transportation Systems (ITSs) as they facilitate information sharing, which improves road safety, traffic efficiency, and provides passengers’ comfort. Due to the dynamic nature of VANETs, vehicles need to exchange the Cooperative Awareness Messages (CAMs) more frequently to maintain network agility and preserve applications’ performance. However, in many situations, broadcasting at a high rate leads to congest the communication channel, rendering VANET unreliable. Existing broadcasting schemes designed for VANET use partial context variables to control the broadcasting rate. Additionally, CAMs uncertainty, which is context-dependent has been neglected and a predefined fixed certainty threshold has been used instead, which is not suitable for the highly dynamic context. Consequently, vehicles disseminate a high rate of unnecessary CAMs which degrades VANET performance. A good broadcasting scheme should accurately determine which and when CAMs are broadcasted. To this end, this study proposes a Context-Aware Adaptive Cooperative Awareness Messages Broadcasting Scheme (CA-ABS) using combinations of Adaptive Kalman Filter, Autoregression, and Sequential Deep Learning and Fuzzy inference system. Four context variables have been used to represent the vehicular context, namely, individual driving behaviors, CAMs uncertainty, vehicle density, and traffic flow. Kalman Filter and Autoregression are used to estimate and predict the CAMs messages respectively. The deep learning model has been constructed to estimate the CAMs’ uncertainties which is an important context variable that has been neglected in the previous research. Fuzzy Inference System takes context variables as input and determines an accurate broadcasting threshold and broadcasting interval. Extensive simulations have been conducted to evaluate the proposed scheme. Results show that the proposed scheme improves the CAMs delivery ratio and decreases the CAMs prediction errors.

To provide dynamic resource management, live virtual machine migration is used to move a virtual machine from one host to another. However, virtual machine migration poses challenges to cloud intrusion detection systems because movement of VMs from one host to another makes it difficult to create a consistent normal profile for anomaly detection. Hence, there is a need to provide an adaptive anomaly detection system capable of adapting to changes that occur in the cloud data during VM migration. To achieve this, the authors proposed a scheme for adaptive IDS for Cloud computing. The proposed adaptive scheme is comprised of four components: an ant colony optimization-based feature selection component, a statistical time series change point detection component, adaptive classification, and model update component, and a detection component. The proposed adaptive scheme was evaluated using simulated datasets collected from vSphere and performance comparison shows improved performance over existing techniques.

With the emergence of WSNs in the recent times, providing trustworthy and reliable data delivery is challenging task due to unique characteristics and constraints of nodes. Malicious node can easily disrupt the integrity of network through the inclusion of false and malicious data and initiate internal attacks. Detection of malicious nodes using trust-based security is an effective and lightweight countermeasure as compared to key based security schemes which incurs higher overhead costs. The WSNs will play greater role in the next-generation IoT systems and a compromised node can jeopardize the availability and authenticity of sensory layer. In this paper, an efficient Belief based trust evaluation mechanism (BTEM) is proposed which isolates the malicious node from trust-worthy nodes and defend against Bad-mouth, On–Off and Denial of Service (DoS) attacks. Bayesian estimation approach is used in gathering direct and In-direct trust values of the sensor nodes which further considers the correlation of the data collected over the time and then estimate imprecise knowledge in decision making for secure delivery of data thus avoiding the malicious nodes. Compared with existing approaches, the proposed BTEM performs better in the detection of malicious node (MN), with lesser delay and improved network throughput.

Online news articles are an important source of information for decisions makers to understand the causal relation of events that happened. However, understanding the causality of an event or between events by traditional machine learning-based techniques from natural language text is a challenging task due to the complexity of the language to be comprehended by the machines. In this study, the Knowledge-oriented convolutional neural network (K-CNN) technique is used to extract the causal relation from online news articles related to the South China Sea (SCS) dispute. The proposed K-CNN model contains a Knowledge-oriented channel that can capture the causal phrases of causal relationships. A Data-oriented channel that captures the position information was added to the K-CNN model in this phase. The online news articles were collected from the national news agency and then the sentences which contain relation such as causal, message-topic, and product-producer were extracted. Then, the extracted sentences were annotated and converted into lower form and base form followed by transformed into the vector by looking up the word embedding table. A word filter that contains causal keywords was generated and a K-CNN model was developed, trained, and tested using the collected data. Finally, different architectures of the K-CNN model were compared to find out the most suitable architecture for this study. From the study, it was found out that the most suitable architecture was the K-CNN model with a Knowledge-oriented channel and a Data-oriented channel with average pooling. This shows that the linguistic clues and the position features can improve the performance in extracting the causal relation from the SCS online news articles. Keywords-component; Convolutional Neural Network, Causal Relation Extraction, South China Sea.

Wireless Sensor Network consists of hundreds to thousands of tiny sensor nodes deployed in the large field of the target phenomenon. Sensor nodes have advantages for its size, multifunctional, and inexpensive features; unfortunately, the resources are limited in terms of memory, computational, and in energy, especially. Network transmission between nodes and base station (BS) needs to be carefully designed to prolong the network life cycle. As the data transmission is energy consuming compared to data processing, designing sensor nodes into hierarchical network architecture is preferable because it can limit the network transmission. LEACH is one of the hierarchical network protocols known for simple and energy saving protocols. There are lots of modification made since LEACH was introduced for more energy efficient purposed. In this paper, hybridization of LEACH-C and LEACH-R and the modification have been presented for a more energy saving LEACH called LEACH-CR. Experimental result was compared with previous LEACH variant and showed to has advantages over the existing LEACH protocols in terms of energy consumption, dead/alive nodes, and the packet sent to Base Station. The result reflects that the consideration made for residual energy to select the cluster head and proximity transmission lead to a better energy consumption in the network.

South China Sea (SCS) generates huge economic value in fishing and shipping lane as well as a high amount of natural resources. Due to its strategic location and high revenue generated, SCS became a place where several nearby countries competed for its territorial claims. Famous territorial disputes such as Spratly islands, Paracel island, Scarborough Shoal happened due to claim on SCS wealth. Newspapers are the main medium that disseminate the message to the public and update whenever SCS conflict happens. News related to SCS events or conflicts usually contain causal relationships between cause and effect. This causal relationship can be extracted and analyzed to obtain the trends of events and conflicts that have happened. In order to avoid any inevitable conflict among countries in SCS region, event prediction is important as it gives a better insight and foresee future events that might happen. In this paper, phrase similarity is used as important metrics for prediction models. First, it extracts news articles based on causality connectors such as "because", "after", "lead to", etc. into tuple. Then, three different embedding techniques, Doc2vec, InferSent and BERT were evaluated based on their best similarity score. The selected embedding technique is used to construct the prediction model and predict South China Sea conflict related events. A crude prediction is done based on similarity of past causes. The result shows that BERT has the highest average accuracy of 50.85% in getting the most similar phrase. By using the causal prediction model, a future possible event can be predicted and this helps to increase the awareness of national security among SCS nearby countries.

Social media has changed the world and play an important role in people lives. Social media platforms like Twitter, Facebook and YouTube create a new dimension of communication by providing channels to express and exchange ideas freely. Although the evolution brings numerous benefits, the dynamic environment and the allowable of anonymous posts could expose the uglier side of humanity. Irresponsible people would abuse the freedom of speech by aggressively express opinion or idea that incites hatred. This study performs hate speech and offensive language detection. The problem of this task is there is no clear boundary between hate speech and offensive language. In this study, a selected new features set is proposed for detecting hate speech and offensive language. Using Twitter dataset, the experiments are performed by considering the combination of word n-gram and enhanced syntactic n-gram. To reduce the feature set, filter-embedded combining feature selection is used. The experimental results indicate that the combination of word n-gram and enhanced syntactic n-gram with feature selection to classify the data into three classes: hate speech, offensive language or neither could give good performance. The result reaches 91% for accuracy and the averages of precision, recall and F1.

Concept drift is a common issue in data stream mining algorithms that causes prediction models to lose its original performance gradually or abruptly due to the non-stationarity of the data distribution and decision boundaries. To combat concept drifts, prediction models need to be updated periodically or when concept drifts occur to adapt to the current concept. Unfortunately, training deep neural networks often require a large amount of data samples and high computational resource consumption, making adaptation slow when concept drifts occur. This paper proposes an approach by searching for an optimum initial parameter that could be adapted quickly to all possible concept drift situations. The initial parameter search is based on the Reptile [1] algorithm, which had been successfully applied in image classification, which allows a neural network model to learn from a few samples and minimal gradient steps. We argue that using an optimum initial parameter allows prior information to be embedded and makes the prediction model less reliant on training exclusively from new data when concept drift occurs. Experimental results show that this approach performs at least as well as current data streaming algorithms but with the lowest computational overhead.

This study investigates the performance of machine learning algorithms for sentiment analysis of students’ opinions on programming assessment. Previous researches show that Support Vector Machines (SVM) performs the best among all techniques, followed by Naïve Bayes (NB) in sentiment analysis. This study proposes a framework for classifying sentiments, as positive or negative using NB algorithm and Lexicon-based approach on small data set. The performance of NB algorithm was evaluated using SVM. NB and SVM conquer the Lexicon-based approach opinion lexicon technique in terms of accuracy in the specific area for which it is trained. The Lexicon-based technique, on the other hand, avoids difficult steps needed to train the classifier. Data was analyzed from 75 first year undergraduate students in School of Computing, Universiti Teknologi Malaysia taking programming subject. The student’s sentiments were gathered based on their opinions for the zero-score policy for unsuccessful compilation of program during skill-based test. The result of the study reveals that the students tend to have negative sentiments on programming assessment as it gives them scary emotions. The experimental result of applying NB algorithm yields a prediction accuracy of 85% which outperform both the SVM with 70% and Lexicon-based approach with 60% accuracy. The result shows that NB works better than SVM and Lexicon-based approach on small dataset.

Most of Vehicular Ad Hoc Network’s (VANET) and Fly Ad Hoc Network (FANET) applications, protocols, and service rely on cooperativeness among network participants. Vehicles, could be unmanned aerial vehicles (e.g. drones), share their sensors’ information to improve network performance and provide safety, traffic efficiency, and entertainments. The accuracy and reliability of this information are vital for VANET performance. However, misbehaving (or faulty) vehicles, which exploit the cooperativeness feature by sharing bogus information, can disrupt the fundamental operations of any potential application and cause loss of people lives and properties. Unfortunately, existing detection approaches cannot effectively thwart such attacks. The highly dynamic vehicular context has been vastly neglected as well as they rely on predefined and static security perimeters to differentiate between false and correct information. In this paper, a context-aware detection approach is proposed to improve the detection accuracy and reduce the false alarm rate. Firstly, the features that well represent the vehicular context has been extracted. Then, an online unsupervised learning method namely the hierarchical clustering algorithm is used to classify the received messages into either genuine or spurious. Finally, Bayesian-based hypothesis testing has been used to confirm the validity of the classification. Results show that the proposed detection approach is promising in effectively detect bogus information attack and improve applications performance.

The emergence and rapid growth of Internet of Things (IoT) with unlimited benefits, facilities and applications provided such as smart cities, smart home, intelligent transportation (ITS), smart health and smart grids impacts everyone's lives. However, IoT based systems and applications are vulnerable to various security threats and attacks due to their deployment and use of sensing devices. Moreover, the lack of standardization due to heterogeneity of devices and technologies implementing security in IoT is real challenge. The aim of this review paper is to highlight the various security threats, challenges and attacks faced by IoT enable applications.

Arabic text sentiment analysis suffers from low accuracy due to Arabic-specific challenges (e.g., limited resources, morphological complexity, and dialects) and general linguistic issues (e.g., fuzziness, implicit sentiment, sarcasm, and spam). The limited resources problem requires efforts to build new and improved Arabic corpora and lexica. We propose a class-specific sentiment analysis (CLASENTI) framework. The framework includes a new annotation approach to build multi-faceted Arabic corpus and lexicon allowing for simultaneous annotation of different facets, including domains, dialects, linguistic issues, and polarity strengths. Each of these facets has multiple classes (e.g., the nine classes representing dialects found in the Arab world). The new corpus and lexicon annotations facilitate the development of new class-specific classification models and polarity strength calculation. For the new sentiment classification models, we propose a hybrid model combining corpus-based and lexicon-based models. The corpus-based model has two interrelated phases to build; (1) full-corpus classification models for all facets; and (2) class-specific models trained on filtered subsets of the corpus according to the performances of the full-corpus models. To calculate polarity strengths, the lexicon-based model filters the annotated lexicon based on the specific classes of the domain and dialect. As a case study, we collect and annotate 15274 reviews from various sources, including surveys, Facebook comments, and Twitter posts, pertaining to governmental services. In addition, we develop a new web-based application to apply the proposed framework on the case study. CLASENTI framework reaches up to 95% accuracy and 93% F1-Score surpassing the best-known sentiment classifiers implemented in Scikit-learn library that achieve 82% accuracy and 81% F1-Score for Arabic when tested on the same dataset.

During the last two decades, there has been a tremendous growth in the use of MANETs, not only due to the development of the technology but also due to their high flexibility. MANETs have challenges and limitations and among the major challenge is the routing process because of high dynamic topology and distributed nature. This is the main reason for quick depletion of network resources. Thus, there is a need to develop a routing protocol to fulfill various application requirements and enhance routing paths according to the topology change. In this paper, two protocols are proposed in order to optimally utilize the constrained network resources and reliably detect high-quality links. They are Hybrid Geo-cast Routing (HGR) protocol and Prediction Residual Lifetime of the Link mechanism (PRLL). The optimal and adaptive HGR protocol utilizes geographical location information to limit search area during route discovery process by including only promising search paths to minimize control overhead. Meanwhile, adaptive PRLL uses signal strength to predicate residual lifetime of the link in order to minimize link breakage, error control packets, and reduce packet drop and end to end delay. The experimental results show that the proposed HGR protocol is able to reduce up to 46.67% control overhead. Meanwhile, the proposed adaptive PRLL protocol improved packet delivery ratio by 26%.

Texture analysis is devised to address the weakness of color-based image segmentation models by considering the statistical and spatial relations among the group of neighbor pixels in the image instead of relying on color information of individual pixels solely. Due to decent performance of the gray-level co-occurrence matrix (GLCM) in texture analysis of natural objects, this study employs this technique to analyze the human skin texture characteristics. The main goal of this study is to investigate the impact of major GLCM parameters including quantization level, displacement magnitudes, displacement direction and GLCM features on skin segmentation and classification performance. Each of these parameters has been assessed and optimized using an exhaustive supervised search from a fairly large initial feature space. Three supervised classifiers including Random Forest, Support Vector Machine and Multilayer Perceptron have been employed to evaluate the performance of the feature space subsets. Evaluation results using Edith Cowan University (ECU) dataset showed that the proposed texture-assisted skin detection model outperformed pixelwise skin detection by significant margin. The proposed method generates an F-score of 91.98, which is satisfactory, considering the challenging scenario in ECU dataset. Comparison of the proposed texture-assisted skin detection model with some state-of-the-art skin detection models indicates high accuracy and F-score of the proposed model. The findings of this study can be used in various disciplines, such as face recognition, skin disorder and lesion recognition, and nudity detection.

Color space conversion utilized by many researchers in order to enhance skin recognition performance by projecting the skin color cluster to a more distinctive distribution. In spite of the substantial research effort in this area, finding a suitable color space for face and skin recognition is still an unsolved issue. Deviation of skin tone under different lighting condition, dissimilarity of skin color among different ethnics and races, various camera sensors characteristics, presence of skin-like color objects in image background and variation of skin color tone among different body limbs are among the major challenges in skin recognition. Majority of these challenges are expected to be mitigated through color space conversion. This paper proposes a new hybrid color space by applying Principal Component Analysis technique to skin color cluster in ten existing conventional color spaces including RGB, YCbCr , YUV, nRGB, i1i2i3, YIQ, XYZ, YPbPr , YES, YCgCr . The proposed hybrid color space which termed P1P2P3 consist of the three major Principal Components of these conventional color spaces components. Using Algebraic simplification these principal component has been reformulated in terms of RGB color space. Parametric pixel wised skin detection techniques have been employed in order to evaluate the proposed color space effect on skin detection performance. Three popular supervised classifiers including Multilayer Perceptron, Support Vector Machine and Random Forest has been employed to generate a parametric model of skin color cluster using the proposed color space. Experiment results shows the proposed hybrid color space P1P2P3 with F -score and False Positive Rate 0.960 and 0.041 respectively performed better than the existing conventional color spaces in term of pixel wised skin recognition.

In general, the current text stemmers for Malay texts were not developed for text stemming against social media texts. Therefore, there is a need to develop an enhanced text stemmer that is able to map morphological variants based on the characteristics of non-standard derived word patterns on social media platforms. It deals with noncompliance word patterns (also called noisy texts or micro text) such as misspelled word and texting language which are often being used as informal conversation. This paper proposes an enhanced text stemmer to perform text stemming against social media texts. The investigation focuses on different patterns of non-standard, non-derived words (mechanics, non-standard word formation, code-switching, and slang words) and also non-standard derived words. The experimental results show that the performance of the proposed text stemmer depends on how much “noise” is in social media texts.

Word stemmer (or text stemmer) is used to remove bound morphemes from derived words so that various morphological variants are mapped into common base forms. It is usually used as one of the preprocessing tools in text classification, text mining, and information retrieval tasks. Therefore, the design of an effective text stemmer is crucial for ensuring text stemming process maps morphological variants into correct base forms. This paper investigates the design consideration of an effective text stemmer from the perspective of the Malay language. These design considerations are based on current challenges faced by previous researchers in performing text stemming against Malay texts. By adopting these considerations, an effective text stemmer is expected to address common stemming errors and also, expected to produce promising stemming accuracy.

Analyzing text especially in Malware domain is quite challenging. Even with Natural Language Processing approach, it limits with the absence of specific Named Entity recognizer to extract related entities of malware from unstructured data like text. It is essential to automate the process of extracting information such as Ransomware entity from text and the information extracted could be used as knowledge reasoning like profiling the behaviour of Ransomware using the information available on the Internet. Although the text itself is unstructured, informal text like Internet forum has its problems and challenges to perform the analysis and extraction process. Thus, the performance of machine learning in carrying out the classification of entities from this type of text depending on the complexity of the model. Therefore, this paper presents the comparison of few supervised learning techniques (CRF, Naive Bayes, and SVM) for model training in extracting Ransomware entities from unstructured text in terms of their performance.

Public sector and private organizations began using cybersecurity control in order to defend their assets against cybercriminals attack. Cybersecurity audits assist organizations to deal with cyber threats, cybercriminals, and cyber-attacks thatare growing in an aggressive cyber landscape. However, cyber-attacks and threats become more increase and complex in complicated cyber landscapes challenge auditors to perform an effective cybersecurity audit. This current situation puts in evidens ce the critical need for a new approach in the cybersecurity audit execution. This study reviews an alternative method in the execution of cybersecurity security checks. The analysis is on the character and behavioral of cyber-attacks and threats using feature extraction and selection method to get crucial elements from the common group of cyber-attacks and threats. Cyber-attacks and threats profile are systematic approaches driven by a clear understanding of the form of cyber-attacks and threats character and behavior patterns in cybersecurity requirements. As a result, this study proposes cyber-attacks and threats profiling for cybersecurity audit as a set of control elements that are harmonized with audit components that drive audits based on cyber threats.

In the process of cloud forensic investigation, roles and responsibilities of cloud consumers and cloud providers do not have clear delineation. Nevertheless, Consumers are responsible for collecting and analyzing data from their adopted cloud services for forensic purposes, thus, should have processes in place to identify, prioritize and collect data from cloud components that they are responsible for. Therefore, a cloud forensic process model that would guide a consumer before, during and after when an incident occurs is of vital importance. To develop such model authors have primarily defined requirements that should be met by forensic process models that aim to lead investigations in the cloud. As a result, a cloud forensic process model that takes consumer perspective into account has been proposed. Subsequently, its fulfillment of the proposed requirements is discussed.

Nowadays, the advancement of the sensor technology, has introduced the smart living community where the sensor is communicating with each other or to other entities. This has introduced the new term called internet-of-things (IoT). The data collected from sensor nodes will be analyzed at the endpoint called based station or sink for decision making. Unfortunately, accurate data is not usually accurate and reliable which will affect the decision making at the base station. There are many reasons constituted to the inaccurate and unreliable data like the malicious attack, harsh environment as well as the sensor node failure itself. In a worse case scenario, the node failure will also lead to the dysfunctional of the entire network. Therefore, in this paper, an unsupervised one-class SVM (OCSVM) is used to build the anomaly detection schemes in recourse constraint Wireless Sensor Networks (WSNs). Distributed network topology will be used to minimize the data communication in the network which can prolong the network lifetime. Meanwhile, the dimension reduction has been providing the lightweight of the anomaly detection schemes. In this paper Distributed Centered Hyperellipsoidal Support Vector Machine (DCESVM-DR) anomaly detection schemes is proposed to provide the efficiency and effectiveness of the anomaly detection schemes.

During the last two decades, there has been a tremendous growth in the use of MANETs, not only due to the development of the technology but also due to their high flexibility. MANETs have challenges and limitations and among the major challenge is the routing process because of high dynamic topology and distributed nature. This is the main reason for quick depletion of network resources. Thus, there is a need to develop a routing protocol to fulfill various application requirements and enhance routing paths according to the topology change. In this paper, two protocols are proposed in order to optimally utilize the constrained network resources and reliably detect high-quality links. They are Hybrid Geo-cast Routing (HGR) protocol and Signal Strength and Congestion Avoidance protocol (SSCA). The optimal and adaptive HGR protocol utilizes geographical location information to limit search area during route discovery process by including only promising search paths to minimize control overhead. Meanwhile, adaptive SSCA uses SSCA mechanism in order to enhance link and node quality detection and reduce packet drop. The experimental results show that the proposed HGR protocol is able to reduce up to 46.67% control overhead. Meanwhile, the proposed adaptive SSCA protocol reduces packet drop to 71.20%.

View references (29) Word stemmer is a preprocessing component that has been widely used in many artificial intelligence applications for reducing derived words to their respective root words. There are many existing Malay word stemmers that have been developed to stem affixation words by using various word stemming methods. Therefore, this paper describes the research trends of the existing Malay word stemmers based on morphological structures of Malay language, general word stemming methods and adopted word stemming methods used in the existing word stemmers. Hence, this paper serves as a preliminary reference for improving the existing Malay word stemmers.

In cloud computing datacenter, task execution delay is a common phenomenal cause by task imbalance across virtual machines (VMs). In recent times, a number of artificial intelligence scheduling techniques are applied to reduced task execution delay. These techniques have contributed toward the need for an ideal solution. The objective of this study is to optimize task scheduling based on proposed orthogonal Taguchi-based cat swarm optimization (OTB-CSO) in order to reduce total task execution delay. In our proposed algorithm, Taguchi orthogonal approach was incorporated into tracing mode of CSO to scheduled tasks on VMs with minimum execution time. CloudSim tool was used to implement the proposed algorithm where the impact of the algorithm was checked with 5, 10 and 20 VMs besides input tasks and evaluated based on makespan and degree of imbalance metrics. Experimental results showed that for 20 VMs used, our proposed OTB-CSO was able to minimize makespan of the total tasks scheduled across VMs with 42.86, 34.57 and 2.58% improvement over minimum and maximum job first (Min–Max), particle swarm optimization with linear descending inertia weight (PSO-LDIW) and hybrid PSO with simulated annealing (HPSO-SA) and likewise returned degree of imbalance with 70.03, 62.83 and 35.68% improvement over existing algorithms. Results obtained showed that OTB-CSO is effective to optimize task scheduling and improve overall cloud computing performance through minimizing task execution delay while ensuring better system utilization.

Understanding what people think about an idea or how they evaluate a product, a service or a policy is important for individuals, companies and governments. Sentiment analysis is the process of automatically identifying opinions expressed in text on certain subjects. The accuracy of sentiment analysis has a direct effect on decision making in both business and government. Working with the Arabic language is very important because of the growing number of online contents in Arabic and the existing resources are limited and the accuracy of existing methods is low. In this study, we do a survey to highlight Arabic sentiment analysis challenging issues based on two main perspectives: Arabic-specific and general linguistic issues. The Arabic-specific challenges are mainly caused by Arabic morphological complexity, limited resources and dialects, while the general linguistic issues include polarity fuzziness, polarity strength, implicit sentiment, sarcasm, spam, review quality and domain dependence.

Keypoint descriptor plays a significant role in a huge number of computer vision applications. A large amount of effort and a number of techniques are proposed in the literature which tried to build an image patch descriptor in different binary and n-n-binary spaces. Despite considerable performance of some existing techniques, there are still open problems to be resolved such as lack of enough reliability and robustness against some image distortions and transformations, especially brightness change, blur and JPEG compression. To address these issues, a keypoint descriptor which is adapted from Tomographic Image Reconstruction is proposed in this research. Convolution of predefined Gaussian smoothed sensitivity maps and associated image patch produce a matrix whose entities indicate the average intensity of the pixels at the convolved pixels in the image patch. The initial descriptor is constructed by finding the absolute differences of all possible pairs of matrix. Genetic Algorithm (GA) and Principal Component Analysis (PCA) are used to optimize this descriptor vector to its most discriminative features. Experimental result shows that the proposed descriptor outperformed some existing techniques particularly in brightness change, JPEG compression and blur while it has reasonable performance in other transformations.

The increment of computer technology use and the continued growth of companies have enabled most financial transactions to be performed through the electronic commerce systems, such as using the credit card system, telecommunication system, healthcare insurance system, etc. Unfortunately, these systems are used by both legitimate users and fraudsters. In addition, fraudsters utilized different approaches to breach the electronic commerce systems. Fraud prevention systems (FPSs) are insufficient to provide adequate security to the electronic commerce systems. However, the collaboration of FDSs with FPSs might be effective to secure electronic commerce systems. Nevertheless, there are issues and challenges that hinder the performance of FDSs, such as concept drift, supports real time detection, skewed distribution, large amount of data etc. This survey paper aims to provide a systematic and comprehensive overview of these issues and challenges that obstruct the performance of FDSs. We have selected five electronic commerce systems; which are credit card, telecommunication, healthcare insurance, automobile insurance and online auction. The prevalent fraud types in those E-commerce systems are introduced closely. Further, state-of-the-art FDSs approaches in selected E-commerce systems are systematically introduced. Then a brief discussion on potential research trends in the near future and conclusion are presented.

Feature selection determines the most significant features for a given task while rejecting the noisy, irrelevant and redundant features of the dataset that might mislead the classifier. Besides, the technique diminishes the dimensionality of the attribute of the dataset, thus reducing computation time and improving prediction performance. This paper aims to perform a feature selection for classification more accurately with an optimal features subset using Multivariate Adaptive Regression Splines (MARS) in Spline Model (SM) classifier. A comparative study of prediction performance was conducted with other classifiers including Decision Tree (DT), Neural Network (NN) and Support Vector Machine (SVM) with similar optimal feature subset produced by MARS. From the results, the MARS technique demonstrated the features reduction up to 87.76% and improved the classification accuracy. Based on the comparative analysis conducted, the Spline classifier shows better performance by achieving the highest accuracy (97.44%) compared to other classifiers.

Various methods have been implemented to detect and mitigate malware. Ransomware is one of the rising malware which getting attention from world due to its impact of attack in the cyber space. Detection of potential features of Malware using traditional approach and usage of text mining is nothing new. However, identifying the Ransomware related entity from external sources and unstructured textual data like forum is new exposure towards the application of text mining in malware domain. Therefore, in this paper, a conceptual scheme is proposed to construct a Background Knowledge of Ransomware which necessary to improve the accuracy of NER when classifying the Ransomware related entity from unstructured data like online forum. From this work, the analysis related to malware also could be understood by people who have no or less expertise in Malware domain since it uses the casual text representation that are obtain from user-generated content made publicly.

Text stemmer is one of useful language preprocessing tools in the field of information retrieval, text mining and natural language processing. It is used to map morphological variants of words into base forms. Most of the current text stemmers for the Malay language focused on removing affixes, clitics, and particles from affixation words. However, these stemmers still suffered from stemming errors due to insufficiently address the root cause of these stemming errors. This paper investigates the root cause of stemming errors and proposes stemming technique to address possible stemming errors. The proposed text stemmer uses affixes removal method and multiple dictionary lookup to address various root causes of stemming errors. The experimental results showed promising stemming accuracy in reducing various possible stemming errors.

Cloud computing is becoming an increasing popular technology where clients can subscribe for resources on a pay per use basis. Distributed Denial of Service (DDoS) is a common form of attack that targets cloud computing. Also, new features of cloud computing such as virtualization and virtual machine migration introduces additional challenges to cloud security as studies reports that the detection accuracy of anomaly detection system is degraded during the migration of virtual machine from one host to another. Therefore, providing effective security is paramount to the quality of service in cloud computing. DDoS attack is carried out by compromising vulnerable host which are used to carry out coordinated attack against a target. This paper therefore proposed a model for an adaptive and distributed IDS for cloud computing capable of effectively detecting coordinated attacks and also addressing challenges introduced by migration of virtual machines in cloud IDS.

loud computing requires more reliable, efficient and scalable load balancing algorithm to survive. As one of the main challenges in cloud computing, load balancing facilitate dynamic workload across multiple nodes ensuring that no single node get overloaded. With proper load balancing, resource consumption is maintained at minimum level, enabling scalability, avoiding bottleneck and overprovisioning etc. In this paper, systematic review on existing load balancing techniques currently prevalent in cloud computing was carried out. Load balancing metrics; Response time, Performance, Resource Utilization, Throughput, Cost Overhead, Scalability, Fault Tolerant and Migration Time were used to evaluate the existing techniques. Findings show that the existing techniques mainly focus on reducing response time, completion time, cost and improving throughput. Neither of the techniques was able to unveil efficient load balancing of task scheduling for single and federated cloud environment. However, research such as load balancing of energy consumption, server consolidation, Virtual Machine Migration, are not taken into consideration by the existing techniques. Future research is to unveil efficient multi-objective load balancing of tasks scheduling algorithm with quality of service improvements for homogeneous and federated heterogeneous cloud environment.

A common cloud forensic model proposed by researchers is ‘Cloud-Forensic-as-a-Service’ where consumers have to access it as a service to collect forensic data from cloud environments. The ‘Cloud-Forensic-as-a-Service’ model raises the question of how it collects digital evidence pertaining to an incident which occurred in the cloud. Currently, types of ‘Cloud-Forensic-as-a-Service’ systems in the literature show that the system is controlled and implemented by the cloud provider, where they unilaterally define the type of evidence that can be collected by the system. A serious limitation of this approach is that it does not offer the consumer sufficient means of performing reasonableness checks to verify that the provider is not accidentally or maliciously contaminating the evidence. To address the problem, the paper proposes a conceptual bilateral Cloud-Forensic-as-a-Service model where both consumers and providers can independently collect, verify the equity of the forensic analysis process and try to resolve potential disputes emerging from the independently collected results. The authors have developed a cloud forensic process model to lead common and significant aspects of a bilateral Cloud-Forensics-as-a-Service model. The paper explicitly discusses the concept of a bilateral Cloud-Forensic-as-a-Service model.

Fraudulent websites pose as legitimate sources of information, goods, product and services are propagating and resulted in loss of billions of dollars. Due to several undesirable impacts of Internet fraud and scam, several studies and approaches are focused to identify fraudulent Internet websites, yet none of them managed to offer an efficient solution to suppress these fraudulent activities. With this regard, this research proposes a fraudulent website detection model based on sentiment analysis of the textual contents of a given website, natural language processing and supervised machine learning techniques. The proposed model consists of four primary phases which are data acquisition phase, preprocessing phase, feature extraction phase and classification phase. Crawler is used to obtained data from Internet and data was cleaned to remove non-discriminative noises and reshape into desired format. Later, meaningful and discriminative patterns are extracted. Finally classification phase consists of supervised machine learning techniques to construct the fraudulent website detection model. This research employs 10-fold stratified cross validation technique in order to validate the performance of the proposed model. Experimental results show that the proposed fraudulent website detection model with cross validated accuracy of 97.67% and FPR of 3.49% achieved satisfactory results and served the aim of this research.

Wireless sensor nodes have been deployed in the fields for sensing and monitoring the surrounding environment for the specific purpose. Deployment of sensors in the field is more challenging due to area coverage for monitoring. One of the deployment method is random distribution of sensor nodes, which leads to make various coverage holes in the network and disturb the network data communication. The coverage holes issues are one of the significant research area in wireless senor networks. In literature, mainly three different types of approaches have proposed to address the coverage area issues in the network including distributed, centralized and hybrid approaches. However, these approaches have their own limitations in terms of consuming more energy and utilizing more computational power. In order to overcome these coverage holes in network, we present an intelligent hybrid approach for sensor deployment in the field and enhances the coverage area for sensor deployment. The proposed approach is evaluated in simulation where results show better performance of proposed approach and cover the area and enhance the convergence rate.

The advancement in technology has changed how people work and what software and hardware people use. From conventional personal computer to GPU, hardware technology and capability have dramatically improved so does the operating systems that come along. Unfortunately, current industry practice to compare OS is performed with single perspective. It is either benchmark the hardware level performance or performs penetration testing to check the security features of an OS. This rigid method of benchmarking does not really reflect the true performance of an OS as the performance analysis is not comprehensive and conclusive. To illustrate this deficiency, the study performed hardware level and operational level benchmarking on Windows XP, Windows 7 and Windows 8 and the results indicate that there are instances where Windows XP excels over its newer counterparts. Overall, the research shows Windows 8 is a superior OS in comparison to its predecessors running on the same hardware. Furthermore, the findings also show that the automated benchmarking tools are proved less efficient benchmark systems that run on Windows XP and older OS as they do not support DirectX 11 and other advanced features that the hardware supports. There lies the need to have a unified benchmarking approach to compare other aspects of OS such as user oriented tasks and security parameters to provide a complete comparison. Therefore, this paper is proposing a unified approach for Operating System (OS) comparisons with the help of a Windows OS case study. This unified approach includes comparison of OS from three aspects which are; hardware level, operational level performance and security tests.

Vehicular ad hoc network (VANET) is the key enabler for future intelligent transportation systems' applications. Due to its high mobility, VANETs rely on the availability of accurate and reliable mobility information of the vehicles. However, misbehavior in mobility can lead to catastrophic results in both safety and traffic efficiency. Several drawbacks of existing misbehavior detection models designed for VANETs which impacted the performance of the applications and the security solutions altogether. Machine learning has not been studied extensively in misbehavior detection in VANET. In this paper, an effective misbehavior detection model based on machine learning techniques is proposed. The proposed model consists of four main phases: data acquisition, data sharing, analysis and decision making. New features are derived which represent the misbehavior, environment and communication status in order to effectively detect the misbehavior data. By using Artificial Neural Network (ANN) techniques namely the feed forward and the back propagation algorithms, an effective misbehavior classifier is trained based on historical data that contains both attacker and normal traffic data. A real-world traffic dataset namely NGSIM is used to construct and evaluate the proposed model. Results show significant improvement in the effectiveness of the proposed model in comparison with the existing baseline model.

Availability of accurate and continuous positioning information is a fundamental requirement for Vehicular Ad Hoc Networks. However, existing positioning approaches does not fulfill the required accuracy of many VANET’s applications and services. Integrating GPS with vehicles’ kinematic information (GPS/DR) is widely suggested for vehicular positioning. In many cases where the GPS is unavailable or instable for long time, this integration resulted in inaccurate positioning. Recently, cooperative positioning (CP) based on vehicle-to-vehicle (V2V) communication have been proposed as an alternative for GPS/DR in many ad hoc networks. Even though, CP needs high communication to achieve the required accuracy, which is not guaranteed in VANET harsh environment. In this paper, two-stage integration algorithm is proposed to ensure continuous and accurate positioning information. The proposed algorithm integrates GPS and kinematic sensors measurements, as well as neighboring mobility information based on two cascaded Kalman filters (KF). The first KF is used to integrate GPS information with kinematic sensors measurements (dead reckoning). Meanwhile, the second KF is used to integrate the results of the first KF with the neighboring mobility information. Results show that the proposed algorithm outperformed the conventional integration algorithm in terms of positioning accuracy under the tested scenario.

The increasing violation of Service Level Agreements (SLA) cause as a result of imbalance tasks allocation across Virtual Machines (VMs) has affected consumers’ Quality of Service (QoS) expectations. Researchers in the literature have put forward several models and tried to solve the problem using Artificial Intelligence (AI) scheduling techniques. Significant improvement has been recorded with the need for an ideal solution. In this paper, a multi-objective task scheduling problem with required consumers’ QoS expectations and a scheduling model in relation to the problem is presented. A Dynamic Multi-Objective Orthogonal Taguchi Based-Cat (dMOOTC) algorithm is then proposed to solve the model. CloudSim tool is used for implementation of the proposed algorithm and evaluated with metrics of execution time, execution cost, and QoS. The performance result as compared with Standard Cat Swarm Optimization (CSO), Multi-Objective Particle Swarm Optimization (MOPSO), Enhanced Parallel CSO (EPCSO), Orthogonal Taguchi Based-Cat Algorithm (OTB-CSO) shows the proposed solution outperformed better by returning good consumers’ QoS expectation.

One of the major issues in current routing protocols for mobile ad hoc networks (MANETs) is to develop a routing protocol that can satisfy different applications’ needs and optimize routing paths to cope with protocol topology changes. Thus, developing a new routing protocol to optimize routing paths based on dynamic topology changes is a challenging task in MANET. Position-based routing protocols are more suitable for high node mobility due to their dependability on local position information rather a global network knowledge. However, most existing position-based routing protocols forward packets greedily which leads to inaccurate location information and results in frequent route failures. Furthermore, incorporating fixed-interval beaconing to distribute local positions introduces high overheads. In this paper an on-demand geo-casting routing scheme which makes use of geographical information to establish cost-effective and reliable paths for packets to reach their destination is proposed. Through the proposed scheme, search area during route discovery process includes only promising search paths, which will alleviate the effect of control message overhead. Moreover, each node is able to independently adjust protocol parameters based on network environment, data traffic conditions, and node’s congestion level. The simulation results show that the proposed protocol outperforms existing routing protocols under different network conditions such as various mobility and traffic loads.

In cloud computing, customers-desired Quality of Service (QoS) expectations are quite superficial due to lack of scalable task scheduling solutions that can adjust to long-time changes. Researchers in the literature have put forward several task scheduling algorithms to account for customers' QoS expectations. Unfortunately, most of these algorithms need improvements to ensure the provisioning of better consumers' QoS expectation. In this study, a Multi-Objective QoS model to address customers' expectation based on execution time and execution cost criteria is presented. A Cloud Scalable Multi-Objective Cat Swarm Optimization (CSO) based Simulated Annealing (SA) (CSM-CSOSA) algorithm is then proposed to solve the model. In this method, the Taguchi Orthogonal approach is used to enhanced the SA and incorporated into the local search of the proposed algorithm for enhancing it exploration capability. Implementation of the algorithm is carried out on CloudSim tool and evaluated using one dataset (Normal distributed) and one Parallel Workload (High-Performance Computing Center North(HPC2N)). Quantitative analysis of the algorithm performance is taken based on metrics of execution time, execution cost, QoS and percentage improvement. Result obtained is compared with that of Multi-Objective Genetic Algorithm (MOGA), Multi-Objective Ant Colony (MOSACO) and Multi-Objective Particle Swarm Optimization (MOPSO), where proposed method is able to return substantial performance with improved QoS.

Moving in tandem with internet services nowadays, education institutions should leverage cloud computing which is considered as ready-made platform. One of the key performance index (KPI) for educational institutions progress in teaching and learning process is the level of enhancement in e-learning applications such as adopting the cloud computing. However, trust and risk issues obstruct the adoption process, whereas many questions arise regarding the reliability of cloud as a public platform for running and storing important and confidential data. This paper aims to identify the factors that affect the risk tolerance and user trust in cloud-based e-learning and proposes a method to deal with users and providers requirements. In particular, this paper identify risk tolerance and trust factors in cloud-based e-learning through surveying previous studies the users perspective. Moreover, a case study is included that conducts a survey regarding user trust factors in cloud-based e-learning system in Malaysian Higher Education institutes, which reflects the factors influencing the Malaysian real educational environment. Multi-criteria Decision Making (MCDM) method is used to confirm the most significant user trust factors through Analytical Hierarchical Process (AHP) technique. The outcomes from result analysis show that the identified factors affect user trust in cloud-based e-learning environment significantly.