Computational Science and Technology

In this work, Computed Tomography (CT) brain images are adopted for the annotation of different types of hemorrhages. The ultimate objective is to devise the semantics-based retrieval system for retrieving the images based on the different keywords. The adopted keywords are hemorrhagic slices, intraaxial, subdural and extradural slices. The proposed approach is consisted of three separated annotation processes are proposed which are annotation of hemorrhagic slices, annotation of intra-axial and annotation of subdural and extradural. The dataset with 519 CT images is obtained from two collaborating hospitals. For the classification, support vector machine (SVM) with radial basis function (RBF) kernel is considered. On overall, the classification results from each experiment achieved precision and recall of more than 79%. After the classification, the images will be annotated with the classified keywords together with the obtained decision values. During the retrieval, the relevant images will be retrieved and ranked correspondingly according to the decision values.

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.

In recent years, increased social concern towards hygienic biometric technology has led to a high demand for contactless palm vein biometric. Nonetheless, there are a number of challenges to be addressed in this technology. Among the most important challenges in the hand rotation issue that is caused inadvertently by unrestricted hand posture. In spite of the existing palm ROI region methods, the inadequacies of handling large rotations have never been accounted. In this paper, a rotation-invariant palm ROI detection method is proposed to handle a hand rotation of up to 360º and thus, providing a high flexibility for hand placement on the sensor. Experiments on the benchmark database validate the effectiveness of the proposed contactless palm vein approach.

A behavior is considered abnormal when it is seen as unusual under certain contexts. The definition for abnormal behavior varies depending on situations. For example, people running in a field is considered normal but is deemed abnormal if it takes place in a mall. Similarly, loitering in the alleys, fighting or pushing each other in public areas are considered abnormal under specific circumstances. Abnormal behavior detection is crucial due to the increasing crime rate in the society. If an abnormal behavior can be detected earlier, tragedies can be avoided. In recent years, deep learning has been widely applied in the computer vision field and has acquired great success for human detection. In particular, Convolutional Neural Network (CNN) has shown to have achieved state-of-the-art performance in human detection. In this paper, a CNN-based abnormal behavior detection method is presented. The proposed approach automatically learns the most discriminative characteristics pertaining to human behavior from a large pool of videos containing normal and abnormal behaviors. Since the interpretation for abnormal behavior varies across contexts, extensive experiments have been carried out to assess various conditions and scopes including crowd and single person behavior detection and recognition. The proposed method represents an end-to-end solution to deal with abnormal behavior under different conditions including variations in background, number of subjects (individual, two persons or crowd), and a range of diverse unusual human activities. Experiments on five benchmark datasets validate the performance of the proposed approach.

Wireless sensor networks are applied in various areas like smart grid, environmental monitoring, health care, and security and surveillance. It applies to many fields, but as the utilization is higher, security becomes more important. Recently, the authentication scheme for the environment of wireless sensor network has also been studied. Wu et al. has announced a three-factor user authentication scheme claiming to be resistant to different types of attacks and maintain various security attributes. However, their proposal has several fatal vulnerabilities. First, it is vulnerable to the outsider attack. Second, it is exposed to user impersonation attack. Third, it does not satisfy user anonymity. Therefore, in this paper, we describe these vulnerabilities and prove Wu et al.’s scheme is unsafe.

In digital age, the valuable asset of every company is their data. They contain personal information, companies and industries data, sensitive government communications and a lot of more. With the rapid development in IT technology, accessing the network become cheaper and easier. As a result, organizations are more vulnerable to both insiders and outsider threat. This work proposes user profiling in anomaly detection and analysis of log authorization. This method enables companies to assess each user’s activities and detect slight deviation from their usual pattern. To evaluate this method, we obtained a private dataset from NextLabs Company, and the CERT dataset that is a public dataset. We used random forest for this system and presented the results. The result shows that the algorithm achieved 97.81% of accuracy.

This research proposes an agent-based framework for solving reallife curriculum-based University Course Timetabling problems (CB-UCT) at the Universiti Malaysia Sabah, Labuan International Campus (UMSLIC). Similar to other timetabling problems, CB-UCT in UMSLIC has its own distinctive constraints and features. The proposed framework deal with the problem using a distributed Multi-Agent System (MAS) environment in which a central agent coordinates various IP agents that cooperate by sharing the best part of the solution and direct the IP agents towards more promising search space and hence improve a common global list of the solutions. All agents are incorporated with Integer programming (IP) search methodology, which is used to generate initial solution in this, regards as well. We discuss how sequential IP search methodology can be incorporated into the proposed multi-agent approach in order to conduct parallel search for CB-UCT. The agent-based IP is tested over two real-life datasets, semester 1 session 2016/2017 and semester 2 session 2016/2017. The experimental results show that the agent-based IP is able to improve the solution generated by the sequential counterpart for UMSLIC’s problem instance used in the current study impressively by 12.73% and 17.89% when three and six IP agents are used respectively. Moreover, the experiment also shows that increasing the number of IP agents lead to the better results.

Face recognition is commonly used for biometric security purposes in video surveillance and user authentications. The nature of face exhibits non-linear shapes due to appearance deformations, and face variations presented by facial expressions. Recognizing faces reliably across changes in facial expression has proved to be a more difficult problem leading to low recognition rates in many face recognition experiments. This is mainly due to the tens degree-of-freedom in a non-linear space. Recently, non-linear PCA has been revived as it posed a significant advantage for data representation in high dimensionality space. In this paper, we experimented the use of non-linear kernel approach in 3D face recognition and the results of the recognition rates have shown that the kernel method outperformed the standard PCA.

Augmented Reality is an exciting technology that enables user to view virtual object overlaying physical environment to increase the student-learning outcome. Objectives: To find if visualization plays a vital role mobile learning through comparison between three groups that were analysed using SmartPLS 3 using Measurement and Structural Model, Traditional, Mobile and mobile-Augmented Reality (mAR) Groups. The analysed data were collected at five universities. Methods: Quantitative method was conducted with students using questionnaire that covers Effectiveness, Self-Efficacy, Motivation, Satisfaction and Features. This paper also presents the comparison between the three groups in terms of Effectiveness towards several relationships that were analysed using SmartPLS to find significant relationship between the construct. It was found that Traditional had one significant relationship where as Non-mAR and mAR both have two significant relationships.

Computational optimization via artificial intelligence has been considered as one of the key tools to gain competitiveness in Industrial Revolution 4.0. This paper proposes computational optimization analysis for designing the widely used industrial control systems - feedforward and feedback control schemes. Although several different optimal tunings for servo and regulatory control problems exist, their applications often present some challenges to plant operators. Plant operators often face difficulties to obtain satisfactory PID controller settings by using the conventional tuning methods, which rely heavily on engineering experience and skills. In the proposed intelligent tuning method for the feedforward plus feedback control system, the closed-loop stability region was first established, which then shall provide the upper and lower limits for computational optimization analysis via Genetic Algorithm. Based on a jacketed reactor case study, the performance of feedforward plus feedback control scheme tuned via Genetic Algorithm was compared to that tuned via Ziegler-Nichols tuning. Comparison of performances showed that computational optimization method via Genetic Algorithm gave improved performances in terms of servo and regulatory control objectives.

A Smart Verification Algorithm (SVA) used with Internet of Things (IoT) applications is proposed performing a verification procedure to enable authorized requests by user to access a smart system with help of Quick Response (QR) tag. It uses encrypted QR-tag values to compare them to original values. Three-layers have been proposed for this verification procedure to attain security objectives. The first layer implements a comparison to preserve the system integrated. In the second layer, original values are stored in offline database storage to disable any access caused by threats; to preserve it available. The third one frequently generates an authenticated QR tag using 1-session private key to prevent both information leakage and an unauthorized access if the key was deduced; to keep it confidential. The SVA aims to increase the system privacy. It is evaluated in terms of security factors. Results confirm that it is faster than other competitive techniques. Additionally, results have discussed SVA’s robustness against unauthorized access’s attempts and brute force attack.

The study is to classify human motion data captured by a wrist worn accelerometer. The classification is based on the various daily activities of a normal person. The dataset is obtained from Human Motion Primitives Detection [1]. There is a total of 839 trials from 14 activities performed by 16 volunteers (11 males and 5 females) ages between 19 to 91 years. A wrist worn tri-axial accelerometer was used to accrue the acceleration data of X, Y and Z axis during each trial. For feature extraction, nine statistical parameters together with the energy spectral density and the correlation between the accelerometer readings are employed to extract 63 features from the raw acceleration data. Particle Swarm Organization, Tabu Search and Ranker are applied to rank and select the positive roles for the later classification process. Classification is implemented using Support Vector Machine, k-Nearest Neighbors and Random Forest. From the experimental results, the proposed model achieved the highest correct classification rate of 91.5% from Support Vector Machine with radial basis function kernel.

The quarter-sweep scheme has been used in solving boundary value problems efficiently. In this paper, we aim to determine the capability of the family of Gauss-Seidel iterative methods to solve the Poisson image blending problem, which are the Full-Sweep Gauss-Seidel (FSGS), Half-Sweep Gauss-Seidel (HSGS) and Quarter-Sweep Gauss-Seidel (QSGS). Second order finite difference approximation is used for the discretization of Poisson equation. Finally, the numerical results show that QSGS iterative scheme is more competent as compared with the full- and half-sweep approaches while obtaining the same quality of output images.

This research work explores the possibility of using deep learning to produce an autonomous system for detecting potholes on video to assist in road monitoring and maintenance. Video data of roads was collected using a GoPro camera mounted on a car. Region-based Fully Convolutional Networks (RFCN) was employed to produce the model to detect potholes from images, and validated on the collected videos. The R-FCN model is able to achieve a Mean Average Precision (MAP) of 89% and a True Positive Rate (TPR) of 89% with no false positive.

The current study presents Integer Programming (IP) search methodology approaches for solving Curriculum-Based University Course Timetabling problem (CB-UCT) on real-life problem instances. The problem is applied in University Malaysia Sabah, Labuan International Campus Labuan (UMSLIC). This research involves implementing pure 0-1 IP and further incorporates IP into a distributed Multi-Agent System (MAS) in which a central agent coordinates various cooperative IP agents by sharing the best part of the solutions and direct the IP agents towards more promising search space and hence improve a common global list of the solutions. The objectives are to find applicable solutions and compare the performance of sequential and cooperative IP search methodology implementations for solving real-life CB-UCT in UMSLIC. The results demonstrate both sequential and parallel implementation search methodologies are able to generate and improve the solutions impressively, however, the results clearly show that cooperative search that combines the strength of integer programming outperforms the performance of a standalone counterpart in UMSLIC instances.

The problem with recommending shows/programs on linear TV is the absence of explicit ratings from the user. Unlike video-on-demand and other online media streaming services where explicit ratings can be asked from the user, the linear TV does not support any such option. We have to rely only on the data available from the set top box to generate suitable recommendations for the linear TV viewers. The set top box data typically contains the number of views (frequency) of a particular show by a user as well as the duration of that view. In this paper, we try to leverage the feedback implicitly available from linear TV viewership details to generate explicit ratings, which then can be fed to the existing state-of-the-art recommendation algorithms, in order to provide suitable recommendations to the users. In this work, we assign different weightage to both frequency and duration of each user-show interaction pair, unlike the traditional approach in which either the frequency or the duration is considered individually. Finally, we compare the results of the different recommendation algorithms in order to justify the effectiveness of our proposed approach.

With the inception of connected devices in smart homes, the need for user adaptive and context-aware systems have been increasing steadily. In this paper, we present an adaptive model predictive control (MPC) based controller for cyber-physical home systems (CPHS) environment. The adaptive MPC controller is integrated into the existing Energy Efficient Thermal Comfort Control (EETCC) system that was developed specifically for the experimental smart house, iHouse. The proposed adaptive MPC is designed in a real time manner for temperature reference tracking scenario where it is evaluated and verified in a CPHS simulation using raw environmental data from the iHouse.

Examination timetabling problems is the allocation of exams into feasible slots and rooms subject to a set of constraints. Constraints can be categorized into hard and soft constraints where hard constraints must be satisfied while soft constraints are not necessarily to satisfy but be minimized as much as possible in order to produce a good solution. Generally, UMSLIC produces exam timetable without considering soft constraints. Therefore, this paper proposes the application of two algorithms which are Constraint Programming and Simulated Annealing to produce a better solution. Constraint Programming is used to generate feasible solution while Simulated Annealing is applied to improve the quality of solution. Experiments have been conducted with two datasets and the results show that the proposed algorithm managed to improve the solution regardless the different problem instances.

Modeling and analysis play essential parts in a Cyber-Physical Systems (CPS) development, especially for the system of systems (SoS) in CPS applications. Many of today’s proposed CPS models rely on multiple platforms. However, there are massive reusable components or modules in the different platform. And also, the model had to be modied to meet the new system requirements. Nevertheless, existing time model technologies deal with them, but it leads to a massive time consuming and high resource cost. There are two objectives in this paper. One is to propose a new simple and proximate time model (SPTimo) framework to the practical time model of hybrid system modeling and analysis. Another is to present a time task scheduling algorithm, mix time cost and deadline first (MTCDF) based on computation model in the SPTimo framework. Simulation results demonstrate that the MTCDF algorithm achieves the priority the scheduling of tasks with a time deadline, and match with optimal scheduling in time requirement and time cost.

Software defined networking is upcoming agile networking system for computer and telecommunication. The apartness of data plane and control plane are key points of SDN architecture which enables flexibility and programmability of network. But distributed denial of service attack is the main threat for software defined networking architecture as it can send huge traffic directly to the controller. In this paper, we proposed a hypothetical concept of smart controller placement for SDN architecture which can monitor controller’s health status and provide continuous services during the DDoS attack.

This paper aims to discuss the implementation of Malay Language Speech Recognition (SR) using Hidden Markov Model (HMM) system training for Malay preschool children. The system is developed by implementing the architecture of HMM-based Recognizer with different feature extraction algorithm. The system is trained for 16 Malay words by collecting 704 speech samples (640 for training samples and 64 for testing samples). Data is collected from 20 preschool children aged between five to six years old in real time environments. The paper also describes the process flow to develop the architecture of the system. The experimental results show that the highest overall system performance is 94.86% - Train and 76.04% - Testing which is using MFCC (Mel-Frequency Cepstral Coefficient) with 39 extracted feature vectors (MFCC39).

This paper describes a general framework of Multi-agent system which incorporates the hyper-heuristics search methodology with both Great Deluge and Simulated Annealing acceptance criteria respectively. There are three types of agents introduce in the framework which involve the communication between heuristic agents, cooperative agents and mediator agent. The common goal for each agent is to improve the quality of course timetabling solutions until the best solution is found when the termination condition meets. A preliminary experiment have been conducted towards this approach in university course timetabling problem and the results shows the framework is able to increase the quality of existing solution compared with other meta-heuristics which have been studied in the previous researches.

This research presents the study of stochastic algorithms in one of the limited study in Space Allocation Problem. The domain involves the allocation of students into the available rooms which is known as Hostel Space Allocation Problem. The problem background of this domain which related with hard constraints and soft constraints are discussed and the formal mathematical models of constraints in Universiti Malaysia Sabah Labuan International Campus are presented. The construction of initial solution is handled by Constraint Programming algorithm. Two algorithms mainly Great Deluge with linear and non-linear decay rate and Simulated Annealing with linear reduction are proposed to improve the quality of solution. The experimental results show that Simulated Annealing with linear reduction temperature performs well in this domain.

Activity recognition in smart homes has attracted increasing attention from researchers due to its potential to recognize the occupant’s activities of daily living such as showering, putting away laundry, grooming, etc. Recognizing the activities of daily living can help to support and assist the older adults, and enable them to continue living independently within their own homes. In order to support the occupants, activity recognition algorithms need to learn from a series of observations obtained from sensors. The central question that this paper aims to address is which sensors are informative for activity recognition. In this paper, the sensor selection problem is addressed using minimum-Redundancy Maximum-Relevance (mRMR) method.

The internet has undoubtedly become everywhere essential to majority of the people from business services to entertainment. Early prevention of network service faults can greatly improve customer satisfaction and experience. Proactive network service faults prediction can certainly help the internet service providers to reduce service workloads and costs. One of the assumptions often made by standard supervised learning is to treat each session log (generated from the network management system) as an individual instance, where each instance is assigned a class label. Although such assumption is appropriate in some domains, it may not be appropriate in network service fault prediction since a network service fault is represented by a collection of session logs. In this paper, we aim to improve the network service fault prediction by transforming the single-instance learning to a multi-instance learning problem. We evaluate our proposed method on a real-world network data and compared with the baseline single-instance learning method. The multi-instance learning approach achieves a higher AUROC performance to single-instance learning approach.

The objective of this research is to collect and analyse road surface conditions in Malaysia using Internet-of-Things (IoT) sensors, together with the development of a machine learning model that can identify these conditions. This allows for the facilitation of low cost data acquisition and informed decision making in helping local authorities with repair and resource allocation. The conditions considered in this study include smooth surfaces, uneven surfaces, potholes, speed bumps, and rumble strips. Statistical features such as minimum, maximum, standard deviation, median, average, skewness, and kurtosis are considered, both time and frequency domain forms. Selection of features is performed using Ranker, Greedy Algorithm and Particle Swarm Optimisation (PSO), followed by classification using k-Nearest Neighbour (k-NN), Random Forest (RF), and Support Vector Machine (SVM) with linear and polynomial kernels. The model is able to achieve an accuracy of 99%, underlining the effectiveness of the model to identify these conditions.

This paper proposed a Real-Time Optimal Trajectory Correction (ROTC) algorithm designed to be applied for autonomous omnidirectional robot. It is programmed to work when a robot undergoes a deviation, an admissible trajectory correction path is generated for the robot rapidly returns to the route line. For the algorithm to do this, initially a deviation scheme is employed to sense deviation and formulates a vector consists of displacement and angle. Via the vector, an admissible correction path is originated utilizing Hermite cubic spline method fused with time and tangent transformation schemes. A Dead Reckoning (DR) technique is applied for robot to pursue the path. Several experiments are arranged to evaluate the reliability of robot navigation with and without the algorithm. It motion is mapped in Graphical User Interface (GUI) window using data from Laser Range Finder (LRF) sensors as attached to the robot controller. Using the map, the performances of the algorithm are evaluated in terms of distance travel and duration to return on the line. The results signify robot navigation with the algorithm required shorter distance and duration as compared to robot navigation without ROTC. Thus, it justifies the algorithm is feasible in the navigation system where it can assist robot effectively to move back to the route line after experiencing a deviation caused by a disturbance.

Autonomous spatial layout design had become the prominent applications for early planning process for a space arrangement. The available spatial layout design applications had focused on the structural and functional demands for a trendy, cultural influences and scalable spatial design. The previous research on the autonomous spatial layout design had proved that the Genetic Algorithm (GA) with full-fledged of operators are able to design an optimal spatial layout that is scalable for the space utilization. However, in these recent years, due to many occurrences of emergency incidents, the security assurance of the pedestrian to evacuate from the spatial layout during panic situation had become the main focus in designing a spatial layout. Hence, this research had proposed the pedestrian movement simulation using Cellular Automata (CA) with the Moore Neighborhood transition movement direction as the objective function for optimizing the GA operators based spatial layout design. The results have shown that the CA based pedestrian movement simulation was able to optimize the GA operators based spatial layout design in designing a feasible and scalable space arrangement with low-risk of pedestrian casualties.

This paper proposes a queuing technique for important video frame packets with the objective to improve the performance of video transmission as perceived by the end users, across the IEEE 802.11e network. The proposed mechanism preserves the video Quality of Experience (QoE) by avoiding the I-Frames transmitted as part of the Group of Pictures (GoP) from being dropped during queue congestion. The method is evaluated using the NS-3 simulator with the Evalvid module and the results demonstrate the video flows will have better in Mean Opinion Score from the subjective evaluation point of view compared to the original IEEE 802.11e queueing.

Address resolution (AR) process, one of the important neighbor discovery protocol (NDP) functions, aims to obtain the corresponding relationship between Internet protocol and media access control addresses. This process uses two NDP messages, neighbor solicitation (NS) and neighbor advertisement (NA) messages, which are unsecure by design. In addition, the target address is revealed in the traditional AR process. Thus, any malicious node on the same link can modify the message and launch denial of service (DoS) attacks. The current mechanisms suffer from high-complexity issue or other forms of security issues that can induce DoS attack on AR in IPv6 link-local network. To overcome these limitations, this work proposes AR-match technique to secure AR process by hiding the target address by using a hash function algorithm and adding a new option named AR-match, which is attached to each NS and NA message for them to become NS- and NA-match messages, respectively. AR-match technique can provide a high security with less complexity and will completely prevent DoS attacks during AR in the IPv6 link-local network.

The implementation of the neighbor discovery protocol has introduced new security vulnerabilities to Internet protocol version 6 (IPv6) networks. One of the most common attacks being attributed to the IPv6 network layer is the denial of service (DoS) router advertisement (RA) flooding attack. An attacker can flood massive amounts of RA packets to the IPv6 multicast address which cause the hosts inside the link-local network to run out of central processing unit resources due to packet processing overhead. This research proposes a hybrid approach of entropy-based technique combined with the adaptive threshold algorithm to detect the aforementioned attack. By dynamically adapting the threshold and choosing the right entropy feature, the proposed technique is able to detect various scenarios of DoS RA flooding attack, including evasion techniques used by attackers. The proposed technique yields 98% detection accuracy according to the experiment conducted..

This paper provides a brief overview of the techniques used in frequent itemset mining. It discusses the search strategies used; i.e. depth first vs. breadth-first, and dataset representation; i.e. horizontal vs. vertical representation. In addition, it reviews many techniques used in several algorithms that make frequent itemset mining more efficient. These algorithms are discussed based on the proposed search strategies which include row-enumeration vs. column-enumeration, bottom-up vs. top-down traversal, and a number of new data structures. Finally, the paper reviews on the latest algorithms of colossal frequent itemset/pattern which currently is the most relevant to mining high-dimensional dataset.

This paper presents the verification and validation processes in producing a realistic bipartite network model to detect dengue hotspot in Sarawak. Based on the result of previous published work, ranking of location nodes of possible dengue hotspot at Sarawak are used to illustrate the validation by comparing the Spearman rank correlation coefficients (SRCC) between the network models. UCINET 6 is used to generate a benchmark ranking result for model verification. A centrality measure analysis feature available in UCINET is used to determine the node centrality of a network model. The validation results show strong ranking similarity for all three groups of network models with good Spearman rank correlation coefficients values of 1.000, 0.8000 and 0.8824 (ρ>0.80; p<0.001) respectively. The top-ranked locations are seen as dengue hotspots and this study demonstrate a new approach to model dengue transmission at district-level by locating the hotspots and prioritizing the locations according to vector density.

Computer networks are aimed to be secured from any potential attacks. Intrusion Detection systems (IDS) are a popular software to detect any possible attacks. Among the mechanisms that are used to build accurate IDSs, classification algorithms are extensively used due to their efficiency and auto-learning ability. This paper aims to evaluate classification algorithms for detecting the dangerous and popular IPv6 attacks which are ICMPv6-based DDoS attacks. A comparison between five classification algorithms namely Decision Tree (DT), Support Vector Machine (SVM), Naïve Bayes (NB), K-Nearest Neighbors (KNN) and Neural Networks (NN) were conducted. The comparison was conducted using a publicly available flow-based dataset. The experimental results showed that classifiers have detected most of the included attacks with a range from 73%-85% for the true positive rate. Moreover, KNN classification algorithm has been the fastest algorithm (0.12 seconds) with the best detection accuracy (85.7%) and less false alarms (0.171). However, SVM achieved the lowest detection accuracy (73%) while NN was the slowest algorithm in training the detection model (323 seconds).

Integrating process is applied to many industries however there are very few research done on it. Determining PID settings for the closed-loop control of integrating process is a challenging task due to its inherent characteristic, which is only stable at one equilibrium operating point. This paper highlighted First Order plus Dead Time in representing process and disturbance model. Improvement of relative performance for transient and steady state response is achieved by using feedforward plus feedback control scheme. Moreover, computational optimization analysis was presented for developing a systematic way to design PID controller for the optimal performance of both servo and regulatory control problems. Performance of the controlled process were then compared in term of graphs, performance index and performance indicator. It is proven and concluded that designed PID controller settings by using computational optimization analysis eventually gives the best performance compared to other tuning methods for a Pumped-tank function of LOOP-PRO simulation software.

It has been more than a decade since the Wireless Local Area Networks (WLAN) based on the IEEE 802.11 standard family has become commercialized. Inexpensive WLAN access points have found their ways in almost every household and enterprise and WLAN devices are embedded in the chips of laptops, tablets, mobile phones, printers, and many other household and commercial appliances. In recent years, there has been a tremendous growth in the deployment of IEEE 802.11-based WLANs under the brand name Wireless Fidelity (Wi-Fi). This growth is as a result of low-cost, international standard (e.g. 802.11a, b, g, n, and ac) flexibility and mobility offered by the technology. However, WLANs are deployed on a non-planning basis, not like public cellular phone networks. In WLAN, nodes commonly use omnidirectional antennas to communicate with access point. Omni-directional antennas may not be efficient due to interface caused by the transmission of packets in all the directions. Many researchers have evaluated the performance of Dense Wireless Local Area Network (a saturated network). There is need to evaluate the performance of densely deployed wireless networks (an area where high number of WLANs are deployed). This research paper has evaluated the performance of densely deployed WLANs using directional and omnidirectional antennas. Optimized Network Engineering Tool (OPNET) Modeler 17.1 has used to simulate directional and omnidirectional antennas.

This paper presents a methodology to measure and analyze mass opinion towards underdeveloped forms of art such as independent films using data mining and topic modelling approach, rather than limited sampling through traditional movie revenue and surveys. Independent films are cultural mediums that foster awareness and social transformation through the advocacies and social realities they present. This methodology helps in addressing challenges of film stakeholders and cultural policy-making bodies in assessing cultural significance of independent films. Twitter has allowed innovative methods in data mining to understand trends and patterns that provide valuable support in decision making to domain experts. By determining the status of Philippine Cinema using social media data analytics as the primary source of the collective response, film stakeholders will be provided with a better understanding how the audience currently interprets their films with results as quantitative evidence. We use the tweets from the Pista ng Pelikulang Pilipino given the festival objective of showcasing films that enhances “quality of life, examine the human and social condition, and contribute to the nobility and dignity of the human spirit”.

This research modeled previously unarticulated experience of perception and skill learning in two different learning environments, namely Augmented Reality Learning Environment (ARLE) and Virtual Reality Learning Environment (ViRLE). An analytical literature review, primarily from Augmented Reality (AR) and Virtual Reality (VR) learning environments, human factor approach, user-based experimentation domains, methodological and contextual variations, was carried out to bring the dispersed research evidences together, organize and develop them into a meaningful conceptual structure. Sixty undergraduate participants, who were illiterate in computer subjects were selected based on participants’ background. They provided their consent and actively participated as two equal groups of thirty participants, in both environments. This experiment was guided primarily by cognitive task analysis and user modelling techniques. Data elicited from this experiment included verbal protocols, video recording, observers’ field notes, performance tests and responses to questionnaires. Participants’ implicit mental models, such as Cognitive Model, Artefact Model and Task Model gradually evolved from numerous iterative cycles of re-construction, analyses and refinement from these data.

Solving the unconstrained optimization problems using Newton method will lead to the need to solve linear system. Further, the Explicit Group iteration is one of the numerical methods that has an advantage of the efficient block iterative method for solving any linear system. Thus, in this paper to reduce the cost of solving large linear system, we proposed a combination between Newton method with four-point Explicit Group (4-point EG) block iterative method for solving large scale unconstrained optimization problems where the Hessian of the Newton direction is tridiagonal matrices. For the purpose of comparison, we used combination of Newton method with basic iterative method namely successive-over relaxation (SOR) point iteration and Newton method with two-point Explicit Group (2-point EG) block iterative method as reference method. The proposed method shows that the numerical results were more superior compared to the reference methods in term of execution time and number of iteration.

The paper presents the classification model of detecting depression based on local binary pattern (LBP) texture features. The study used the video recording from the SEMAINE database. The face image is cropped from a video and extracting the Uniformed LBP features in every single frame. Video keyframe extraction technique was applied to improve frame sampling to a video. Using the SVM with RBF kernel on the original ULBP features, result showed an accuracy of 98% on identifying a depressed person from a video. Also, part of the classification is to implement Principal Component Analysis on the original ULBP features to analyze facial signals by comparing both of the accuracy results. Using the original ULBP features with SVM applying radial basis function kernel, it resulted higher in accuracy whereas the result of using only ten features computed from the PCA of the original ULBP features. The result of the PCA decreased by 5% gaining only 93% in accuracy applying the same cost and gamma values of SVM RBF kernel used on the original ULBP features.

Due to the rapid growth of targeted malware attacks, malware analysis and family classification are important for all types of users such as personal, enterprise, and government. Traditional signature-based malware detection and anti-virus systems fail to classify the new variants of unknown malware into their corresponding families. Therefore, we propose malware family classification system for 11 malicious families by extracting their prominent API features from the reports of enhanced and scalable version of cuckoo sandbox. Moreover, the proposed system contributes feature extraction algorithm, feature reduction and representation procedure for identifying and representing the extracted feature attributes. To classify the different types of malicious software Random Forest (RF), K-Nearest Neighbor (KNN), and Decision Table (DT) machine learning multi-class classifiers have been used in this system and RF and KNN classifiers provide 95.8% high accuracy in malware family classification.

Ciphering algorithms play a main role in this digital era, especially when data are transferred via internet. Many algorithms has been developed and used to encrypt and decrypt data for secure storing or transmission. At present, both synchronous and asynchronous encryptions are used to achieve the high security and to speed up the encryption time and process. Advanced Encryption Standard (AES) plays a prominent role among the synchronous encryption algorithms. It is also known as Rijndael Algorithm. Because of high performance of AES algorithm, it has been chosen as a standard among symmetric cipher algorithms. In this paper, we would like to propose a symmetric encryption algorithm. Modification is based on AES and we add an additional or second key. Another modification is also done at SubBytes step by adding the transportation operation in the original SubBytes operation. To analyze the performance of the modified proposed algorithm, Java language is used to implement the algorithm and then the performance is analyzed. After analyzing and verifying the experimental results, the proposed revised algorithm also shows good performance and high security from the cryptographic point of view. Based on the results of comparison between modified AES and original AES algorithm, our proposed algorithm can be used as a symmetric encryption algorithm, especially for the applications that share sensitive data files via insecure network.

This paper focuses on the design of a WiFi-based tracking and monitoring system that can detect people’s movements in a residential neighbourhood. The proposed system uses WiFi access points as scanners that detect signals transmitted by the WiFi-enabled smartphones that are carried by most people. Our proposed system is able to track these people as they move through the neighbourhood. We implement our WiFi-based tracking system in a prototype and demonstrate that it is able to detect all WiFi devices in the vicinity of the scanners. We describe the implementation details of our system as well as discuss some of the results that we obtained.

Bayesian Networks have been widely used in various domains, but they have been rarely used in educational domain. In this paper, we discover a Bayesian Network model to figure out variables related to adolescents’ mobile phone dependence and their influences. For this study, Markov Blanket is used to identify the strongly related variables with the Korea Children and Youth Panel Survey (KCYPS) data. From the analysis with the discovered BN, “attention ability”, “depression”, “caregiver’s abuse”, “fandom activity”, and “aggression” are extracted as the variables related to adolescents’ mobile phone dependence. These results suggest that considering the variables and their interactions are useful to adjust adolescent’s mobile phone dependence. This paper also shows Bayesian Networks are adequate to find the interdependence of variables and their causal relationships in educational domain.

In this research, our focus is to establish a relationship between some of the entrepreneurial best practices such as good networking skills, developing a clear vision, perseverance and ability to take risks with the business success in the field of kiwifruit contractors. Failures at the initial stage of this business is a common occurrence in the Bay of Plenty region of New Zealand. For aspiring kiwifruit contractors achieving success is a herculean but a possible task. The success factor in this research is calculated based on the number of hectares cultivated land and the number of employees hired by the contractors. The research design adopted in this study is the quantitative research approach, the instrument of a well-structured questionnaire was devised, which was based on the 5 point Likert scale format. Weka, a well known data mining toolbox was used for the analysis of primary data collected from the respondents. In this research, rule based and decision tree algorithms were used to extract useful and actionable information from the data. The study concluded that clear vision and risk taking capabilities are two most important features required to become successful in this business.

This paper describes the development of an intelligent irrigation management system that can be used by farmers to manage water allocation in the farms. Each farm is represented as a single agent that can work out the actual water required for each crop in the farm based on the crop’s drought sensitivity, growth stage, the crop coefficient value and the soil type. During water scarcity, this system can prioritise irrigation allocation to different crops on a farm. Our initial experiment showed that using the irrigation management system, the farm can achieve a consistent water reduction which is more than the required reduction. The results showed that the agent consistently recorded water reduction higher than the actual reduction required by the water authority. This significant reduction means that more water can be conserved in the farm and reallocated for other purposes.

Agent technology is a new emerging paradigm for software systems. In order to fully utilize the capability of this technology, multiple agents operate in software environment by cooperating, coordinating or negotiating with each other. However, these interactions require these agents to communicate with each other through a common language or protocol. Agent communication language (ACL) is a vital component in multiagent system (MAS) to enable the agents to communicate and exchange messages and knowledge. However, there are no universally agreed agent communication language that is widely adopted. Different agent communication languages and different semantic models have been developed to ease the communication between agents in MAS. The purpose of this paper is to review and highlight advances in the development of ACL.

Magnetic field detection has been widely accepted in many applications such as military systems, outer space exploration and even in medical diagnosis and treatment. Low magnetic field detection is particularly important in tracking of magnetic markers in digestive tracks or blood vessels. The presence of magnetic fields’ strength and direction can be detected by a device known as magnetometer. A magnetometer that is durable, room temperature operation and having non-movable components is chooses for this project. Traditional magnetometer tends to be bulky that hinders its inclusion into micro-scaled environment. This concern has brought the magnetometer into the trend of device miniaturization. Miniaturized magnetometer is usually fabricated using conventional microfabrication method particularly surface micromachining in which micro structures are built level by level starting from the surface of substrates upwards until completion of final structure. Although the miniaturization of magnetometer has been widely researched and studied, the process however is not. Thus, the process governing the fabrication technique is studied in this paper. Conventional method of fabrication is known as surface micromachining. Besides time consuming, this method requires many consecutive steps in fabrication process and careful alignment of patterns on every layer which increase the complexity. Hence, studies are done to improve time consuming and reliability of the microfabrication process. The objective of this research includes designing micro scale magnetometer and complete device fabrication processes. A micro-scale search coil magnetometer of 15 windings with 600μm thickness of wire and 300μm distance between each wire has been designed.

This paper reviews the recognition-based graphical password system. Twenty-five recognition-based graphical password systems are studied and analyzed with regards to their security threats. Countermeasures and suggestions are given to prevent and reduce the security threats. A comparison summary of the selected recognition-based graphical password system is deliberated at the end of this paper.

Malware threats are a persistent problem that interrupts the regular utilization of IT devices. For effective prevention of malware infections in computer system, development of a malware mitigation system needs to be developed. Malware mitigation system should encompass a thorough technical and management outlook to achieve an effective result. A Management Framework should thus be put in place to facilitate better management and effective outcomes of such a system. This research presents the identification, formulation and proposal of a Management Framework for the development of a malware eradication and remediation system to mitigate cyberattacks. The aim of this research is to construct a Management Framework that allows for the effective development of a malware eradication and remediation system. The method used in this work is qualitative research (observation and interviews) at organizations that have implemented similar systems. The framework covers specific areas that refer to the management of people, process and technology in designing a malware eradication and remediation system.

Unmanned Aerial Vehicle (UAV) is a type of autonomous vehicle for which energy efficient path planning is a crucial issue. The use of UAV has been increased to replace humans in performing risky missions at adversarial environments and thus, the requirement of path planning with efficient energy consumption is necessary. This study analyses all the available path planning algorithms in terms of energy efficiency for a UAV. At the same time, the consideration is also given to the computation time, path length and completeness because UAV must compute a stealthy and minimal path length to save energy. Its range is limited and hence, time spent over a surveyed territory should be minimal, which in turn makes path length always a factor in any algorithm. Also the path must have a realistic trajectory and should be feasible for the UAV.

Sign language translator devices translate hand gestures into text or voice that allow interactive communication between deaf and hearing people without the reliance on human interpreters. The main focus of this work is the development of a wireless wearable device for a sign language translator using an Intel UP Squared (UP2) board. The developed device is consists of a wearable glove-based wearable and a display device using an Intel UP2 board. When hand gestures have been created by a user, the accelerometer and flex sensors in the wearable are able to measure the gestures and conveyed the data to an Arduino Nano microcontroller. The microcontroller translates the gestures into text, and then transmits it wirelessly to the UP2 board, subsequently displays the text on an LCD. In this article, the developed hardware, circuit diagrams as well as the preliminary experimental results are presented, showing the performance of the device, while demonstrating how the Intel UP2 board can be connected to a low-cost Arduino microcontroller wirelessly via Bluetooth communication.

Advanced Persistent Threat (APT) is an attack aim to damage the system’s data from the aspect of confidentiality and integrity. APT attack has several variants of attacks such social engineering techniques via spear phishing, watering hole and whaling. APTGuard exhibits the ability to predict spear phishing URLs accurately using ensemble learning that combines decision tree and neural network. The URL is obtained from the SMS content received on the smart phones and sent to the server for filtering, classifying, logging and finally informing the administrator of the classification outcome. APTGuard can predict and detect APT from spear phishing but it does not have the ability of automated intervention on the user receiving the spear phishing URL. As a result, APTGuard is capable to extract the features of the URL and then classify it accordingly using ensemble learner which combines decision tree and neural network accurately.

Smart home is a home that has a set of electrical appliances connected to a network can be remotely controlled, accessed and monitored. Smart home normally is used to fine-tune human’s daily life. However, the main problem is when different devices such as sensors and electrical appliances produced by different manufacturers being used in a smart home. Each device produced by different manufacturer might come with their own set of communication protocol. Therefore, it is important to have a standardized communication protocol for interconnecting and remotely controlling those devices to ensure data can be transmitted to each other. This project aims to present the Smart Home using MEMS Sensor and Ambient Intelligence (SAHOMASI) where in this paper, a standardized communication protocol is proposed to overcome the above problem. The features of the system includes monitoring elderly’s activities, detect fall and sending email to caregiver upon abnormal activity or fall detected.

Contextual retrieval is a critical component for efficient usage of knowledge hidden behind the data. It is also among the most important factors for user satisfaction. It essentially comprise of two equally important parts – the retrieval mechanism and the knowledge base from which the information is retrieved. Despite the importance, context aware knowledge bases have not received much attention and thereby, limiting the efficiency of precise context aware retrieval. Such knowledge bases would not only contain information that has been efficiently stored but the knowledge contained would be context based. In other words, machines would understand the knowledge and its context rather than just storing data. This would help in efficient and context aware retrieval. The current paper proposes rules and methodologies for construction of such context aware knowledge bases. A case study to demonstrate the application of the methodology and test the efficiency of the proposed methodology has also been presented. The results indicate that knowledge bases built on these principles tend to generate more efficient and better context aware retrieval results.

Correction to: Chapter “A Review on Energy Efficient Path Planning Algorithms for Unmanned Air Vehicles” in: R. Alfred et al. (eds.), Computational Science and Technology , Lecture Notes in Electrical Engineering 481, https://doi.org/10.1007/978-981-13-2622-6_51 The original version of this chapter was inadvertently published with incorrect first author’s name as “Sulaiman Sanjoy Kumar Debnath”. This has been corrected as “Sanjoy Kumar Debnath” in the chapter.