Thursday, September 19, 2019

These articles have been peer-reviewed and accepted for publication in JICT, but are pending final changes, are not yet published and may not appear here in their final order of publication until they are assigned to issues. Therefore, the content conforms to our standards but the presentation (e.g. typesetting and proof-reading) is not necessarily up to the JICT standard. Additionally, titles, authors, abstracts and keywords may change before publication.

Nursuriati Jamil, Izzad Ramli & Norizah Ardi
The pronunciation of Malay vowels is taught by referring to the International Phonetic Association transcription sound chart. The documentation of pronunciation is done by describing and distinguishing the vowel’s sound through impressionistic. In the last five years, several initiatives were done to quantify the standard Malay vowels pronunciation using formant frequencies. However, only one work was found to measure Malay vowels using formant frequencies for Kedah district. Some of the findings, however, were contradicting as the spoken vowels are extracted from a read speech which may not represent the natural dialect. Therefore, this paper investigated the Malay vowels variations of three districts (Perlis, Kelantan, and Terengganu) using spontaneous speeches acquired in natural settings. Eight (8) Malay vowels were collected from local males and females residing in Perlis, Kelantan, and Terengganu. Four formant frequencies (F1-F4) were measured from the vowels extracted from the spontaneous speeches of the locals. Further analysis of the first and second formant show that the Malay vowels of Terengganu and Kelantan has a broader range of formants and are located in approximately the same position in the mouth.However, the vowels of the Perlis dialects are closely positioned in a narrower mouth space. A Malay vowel diagram is also plotted that can be used for future reference in research and educational tool for language learning. Finally, Kelantan and Terengganu dialects are shown to be very similar based on the first and second formants.
Keywords: formant frequencies, Malay dialect, vowel diagram, vowel recognition.

Badamasi Imam Ya’u, Norsaremah Salleh, Azlin Nordin, Ali A. Alwan, Norbik Bashah Idris & Hafiza Abas
Mobile applications and devices play significant roles in boosting a global business, which encompasses various domain such as health, education, bank, transport, to mention a few. They become indispensable tools for everyday activities, thus these applications are developed in a very large quantity with a diverse features and platforms. This brings about new set of difficulties and security challenges. To address this problem and ensure quality and security of these applications, a rigorous and systematic testing especially, using a cloud-based environment is required. Using a systematic mapping study (SMS) method is significant in reporting emprirical studies that address issues cloud-based mobile application testing. This paper presents a total of 23 primary studies that investigate a cloud based mobile application testing and the effect of a Testing as a Service (TaaS). Majority of these studies (56.5%) contribute to literature with a number of framework proposals. We discover that most of the studies (60.9%) focus on Android applications and usually supporting one type of mobile app testing. In addition, despite a plethora of services offered by TaaS, majority of the studies (52.2%) fail to focus on TaaS. The results of this SMS identify a clear gap: 1) lack of general and scalable approaches to support diverse types of mobile apps testing for applications using various platforms and 2) lack of evaluation methods such as case study to validate the proposed approaches.
Keywords: Cloud-based mobile application testing, systematic mapping, testing-as-a-service.

Nurul Atikah Mohd Sharif, Norazura Ahmad, Nazihah Ahmad, Wan Laailatul Hanim Mat Desa, Khaled Mohamed Helmy,  Wei Chern Ang & Ida Zaliza Zainol Abidin
An emergency department (ED) of a hospital is a critical unit that deals with time-sensitive and life-threatening cases. Rapid treatment and accuracy in diagnosis are considered as main characteristics for excellent operational processes in EDs. However, in reality, long waiting time and uncertainty of diagnosis may affect the quality of ED services. This situation could be improved by utilising computing technologies to assist medical professionals in making fast and accurate decisions. This article is devoted to the issues of under-treatment and uncertainty condition of acute asthma cases in ED. A novel approach known as the fuzzy logic principle is used to determine the severity of acute asthma. The fuzzy set theory is embedded into the expert system (ES) to assess the severity of asthma among patients in ED, which known as Fuzzy Rule-based Expert System for Asthma Severity (FRESAS) determination. The proposed fuzzy methodology effectively deals with the fuzziness of the patient information data and subjective judgment of medical practitioners’ level on eight criteria assessed in severity determination. Knowledge acquisition and representation, fuzzification, fuzzy inference engine, and defuzzification are the processes that pass through the FRESAS development which involves expert advice. System evaluation is performed using dataset extracted from ED’s clerking note from one of the hospitals in Northern Peninsular Malaysia. System evaluation demonstrated that the proposed system could work efficiently in determining the severity of acute asthma and offers extension research for other types of disease in ED as well as improves other hybridisation approaches.
Keywords: Emergency department, acute asthma, fuzzy rule-based.

Ahmad Afif Ahmarofi, Razamin Ramli, Norhaslinda Zainal Abidin, Jastini Mohd Jamil & Izwan Nizal Shaharanee
Multilayer perceptron network (MLP) has a better prediction performance compared to other networks since the structure of MLP is suitable for training process in solving prediction problem. However, to the best of our knowledge, there is no rule of thumb in determining the number of hidden nodes within the MLP structure. Researchers normally test with various numbers of hidden nodes to obtain the lowest square error value for optimal prediction result since none of the approaches is claimed as the best practice. Thus, the aim of this paper is to determine the best MLP network by varying the number of hidden nodes of the developed network in predicting the cycle time for producing a new audio product at a production line. The networks are trained and validated through 100 sets of production lots from a selected audio manufacturer. As a result, the 3-2-1 MLP network is the best network based on the lowest square error value compared to the 3-1-1 and 3-3-1 networks. Consequently, the 3-2-1 predicted the best cycle time of 5 seconds to produce a new audio product. Hence, the prediction result can guide production planners to manage the assembly process at the production line smoothly.
Keywords: Artificial neural networks,multilayer perceptron, hidden node, cycle time, production line. 

Zamira Hasanah Zamzuri, Akmalia Shabadin & Siti Zaharah Ishak
Exploring the cause and effect of hazardous events such as traffic accident is vital to the society. Statistical analyses have been a great help in terms of understanding and making inference on the cause-effect analysis and also predicting the occurrence of the accident in the future. One of the issues that could not be handled by the conventional way of statistical modelling is the interrelationships exist between the variables in the data set. With the advent of technology and the wide application of machine learning algorithm, this goal can be achieved through the Bayesian network analysis, in which it is a directed acyclic probabilistic graphical model. By using Hill Climb (HC) and Tabu algorithms, the structure of the data was learnt and their relationship is estimated through the conditional probability based on the Bayes’ theorem. We found that that weather does impact on the accident occurred through the lighting condition and the traffic system. It is also learnt that fatality accidents have a higher likelihood to occur in head-on, turn over and out of control accidents. The use of Bayesian network allows for the probability queries which is very important estimates needed as we want to know what is the risk that we face given the information that we have in hand.
Keywords: Bayesian network, HC algorithm, Tabu algorithm, traffic accidents.

Farhanah Atiqah Norki, Radziah Mohamad & Noraini Ibrahim
Mobile applications are expected to receive context input such as location, speech, and network from different context providers. Since context can be considered as knowledge, a formal method is needed to capture this knowledge. There is less work on ontology model that could be reused to model a new context ontology for Android mobile application. Therefore, this study proposed an ontology specifically for Android mobile application, COCCC, to formalize context knowledge present within it. METHONTOLOGY method was used to create COCCC ontology as it offers intermediate representation in the form of concepts. The concepts from the context ontology were extracted from various resources, sorted and categorized based on the types and functions for standardization purposes. Survey was given to five domain experts for evaluation of COCCC ontology in terms of its usability. Data from these experts were analysed and the results have confirmed that the proposed context ontology is usable to Android mobile application developers.
Keywords: Context ontology, knowledge representation, mobile application, ontology.

Naseer Sanni Ajoge, Azizi Ab Aziz & Shahrul Azmi Mohd Yusof
Anxiety is an aversive motivational state that occurs when an individual perceives threat at events. This condition creates harmful effects for candidates during the interview session. An interviewee overwhelmed in such states deploys worry as resources to cope with the threat hence losing the ability to present self positively for favourable assessments. Most of the digital approaches to assist interviewee in this condition are focused on coaching for verbal and non-verbal cues. The aspect of understanding interviewees’ psychological complexities that influence their behavioural tendencies is lacking in these approaches. As the first step in building an intelligent digital based therapy platform to overcome this issue, this article provides a building block to understanding the interviewee anxiety state by means of a computational model. The model is developed based on the conceptual model derived from generalized anxiety disorder theories. The formal model is evaluated using mathematical analysis to determine possible equilibria state and the simulation results are tested against known cases in the literature. The simulation results show that the degree of threat perceived at events is based on task demand and the resources to cope. Threat is the building block of anxiety through worry which is controlled by one’s personality and inherent trait anxiety. The results conform to established facts in literature. Consequently, this model can serve as the basis to build an integrated interviewee mental state model embedded with self-efficacy and motivation constructs as a holistic approach to support interviewee during simulated training and coaching environment.
Keywords: Anxiety States, Computational Cognitive Modeling, Intelligent Support Agent, Interviewee Behavioural States

Idheba Mohamad Ali O. Swes, Azuraliza Abu Bakar
Feature construction (FC) refers to a process of using the original features to construct new features with better discrimination ability. Particle swarm optimisation (PSO) is considered a powerful search technique and it has had a successful utilisation in FC. However, the application of PSO for feature construction using high dimensional data is considered as a challenge, due to the large search space and high computational cost. Moreover, ineffectivefeatures may be constructed when PSO is applied for the whole features because it may contain noise, irrelevance, and redundancy. Hence the main purpose of this paper is to select the most informative features and constructing new features from the selected features for better classification performance. To do so, feature clustering methods are used to aggregate similar features into clusters, where the data’s dimensionality is lessened by choosing representative features from every cluster to form the final feature subset. The clustering of features has been proven to be accurate in feature selection (FS); however, there is only one current work that studies its utilisation in FC for the classification with some limitations such as using only two binary classes and no much improvement in the accuracy. This paper proposes a cluster based PSO feature construction approach called ClusPSOFC. The Redundancy-Based Feature Clustering (RFC) algorithm is utilised to choose the most informative features from the original data while PSO is used to construct new features from those selected by RFC. Experimental results obtained from utilising six UCI datasets and six high-dimensional data demonstrate that the proposed method had better performance compared to original full features, other PSO based FC methods and the standard genetic programming based feature construction (GPFC). Hence, ClusPSOFC becomes a competitive method for feature construction in classification of high dimensional data.
Keywords: Particle swarm optimisation, feature construction, genetic programming, classification, high- dimensional data

Priya C V & K. S. Angel Viji
In a password-based authentication technique, users are granted permission to access a secure system when the username and password match the database of the system. In the meantime, anyone who enters the correct username and password of a valid user can log in to the secure network. Sometimes the imposter can hack the legal user's secret password, and it is also not difficult to find out the other person's private password. Thus, the existing structure is exceptionally helpless. One way to strengthen the password-based authentication technique is by keystroke dynamic. Therefore in the proposed keystroke dynamics based authentication system, despite the password match, the similarity between the typing pattern of the typed password and password samples in the training database is verified. The timing features of the user's keystroke dynamics are collected to calculate the threshold values. In this paper, a novel algorithm is proposed to authenticate the legal user based on the empirical threshold values. The first step involves the extraction of timing features from the typed password samples. The password training database for each user is constructed using the extracted features. Further, the empirical threshold limits are calculated from the timing features in the database. The second step involves the user authentication using these threshold values. The experimental analyses are carried out in MATLAB simulation, and the results indicate a significant reduction in false rejection rate and false acceptance rate. The proposed method yields very low equal error rate of 0.5% and the authentication accuracy of 99.5%, which are considered as more suitable and efficient for real-time implementation.The proposed method can be a useful resource for identifying the illegal invasion and is valuable in securing the system as a correlative or substitute form of client validation.
Keywords: Authentication, computer security, empirical threshold, feature extraction, keystroke dynamics.

P. Shanthi Saravanan & S. R. Balasundaram
Use of mobile devices enabled with the internet has increased the rapid development of location-based services (LBS). LBS allow users to access useful information such as nearest ATM, temple, and so on. Though users enjoy the convenience of LBS, they are being exposed to the risk of location disclosures which lead to potential abuse of location data. Hence, location privacy protection has recently received considerable attention in LBS. There are numerous techniques have been presented by various researchers to protect the location-context of the users. Location Cloaking is an often used technique to protect the location-contexts. Most of the existing location cloaking algorithms are concerned only with snapshot user locations and cannot effectively prevent the users from the location-dependent attacks when users’ location-contexts are continuously updated. This paper presents a solution to protect users from the location-dependent attacks by improving the existing clique based cloaking algorithm. The main idea is to maintain maximum sized cliques required for location cloaking in an undirected graph. Thus, a qualified clique can be quickly identified and used to generate the cloaked region when a new request arrives. In addition, dummy queries are generated to protect the users from unusual situations. Through maximum sized clique and dummy query generation more numbers of user queries are getting cloaked with a reasonable amount of time, thereby provides better privacy protection while using the LBS applications. The experimental results show that the proposed cloaking algorithm outperformed existing algorithms such as IClique, OptClique and MMBClique in terms of its cloaking success rate and processing time.
Keywords: Location based services, location-dependent attacks, privacy preservation, spatial cloaking

Dewi Yanti Liliana, T. Basaruddin, M. Rahmat Widyanto & Imelda Ika Dian Oriza
Emotion is an important element in an interaction since it conveys human perception and response about an event. Unlike verbal words that can be manipulated, emotion is brief, spontaneously appears and provides more honest information. Emotion consists of several classes of basic emotion which is a human primary emotion that differ one from another; those classes are happy, sad, fear, surprise, disgust, and angry. Meanwhile, a psychologist has developed a set of rules to recognize emotion based on facial expression. This research intends to develop an artificial intelligent model based on psychologist knowledge to recognize emotions by analyzing facial expressions. Moreover, the proposed model defines high-level fuzzy linguistic features of facial components which distinguishes it from existing methods which commonly use low-level image features (e.g. color, intensity, histogram, texture). High-level linguistic features (e.g. eye is open, nose is wrinkle) are better representing human minds than low-level features which are only understandable by machine. The model initially works by detecting facial points to locate important facial components; extracting geometric facial component features which is then applied to the fuzzy facial component inference system and resulting high-level linguistic facial features; the last step is applying high-level linguistic features to the fuzzy emotion inference system which classifies the input image into the respective emotion class based on psychologist rules. Experiments using facial expression dataset gave a high accuracy rate at 98.26% for fuzzy facial component linguistic identification. The proposed model also outperformed other classifiers (Fuzzy C-Means, Fuzzy Inference System, and Support Vector Machine). This intelligent model contributes to various fields, including psychology, health, and education; especially for helping people with emotion disorders (e.g. Alexithymia, Asperger, and Autistic syndromes) to recognize emotions.
Keywords: basic emotion, emotion recognition, facial expression, facial component, fuzzy system, high-level linguistic features

Human Activity Detection and Action Recognition in videos using Convolutional Neural Networks
Jagadeesh B & Chandrashekar M Patil
Human activity recognition from video scenes has become a significant area of research in the field of computer vision applications. Action recognition is one of the most challenging problems in the area of video analysis and finds applications in human-computer interaction, anomalous activity detection, crowd monitoring and patient monitoring, Several approaches have been presented for human activity recognition using machine learning technique. The main aim of the work is to detect and track of human, and classification of actions for two publicly available video database. In this work, a novel approach of feature extraction from video sequence by combining Scale Invariant Feature Transform (SIFT) and optical flow computation is used where shape, gradient and orientation features are also incorporated for robust feature formulation. Tracking of the human in the video is implemented using Gaussian Mixture Model. Convolutional Neural Network based classification approach is used for database training and testing purpose. The activity recognition performance is evaluated for two public datasets namely Weizmann dataset and KTH dataset with action recognition accuracy of 98.43% and 94.96% respectively. Experimental and comparative study shows that proposed approach outperforms when compared with state-of-art techniques.
Keywords: Action recognition, Convolutional Neural Network, Gaussian Mixture Model, Optical flow, SIFT feature extraction.

N. Sathish Kumar & K. Raja Kumar
Nowadays, Underwater Acoustic Sensor Network (UASN) plays a significant role in many application areas like surveillance, security, commercial and industrial application and so on. In UASN routing, the propagation delay and collision are perennial problems due to data transfer from various sensor nodes to the Sink Node (SN) at the same time. In this paper, we propose a Collision Aware Priority Level mechanism based on the Medium Access Control (MAC) protocol (CAPL-MAC) for transferring the data from the Sensor Head (SH) to the SN. In the proposed protocol, we used Parallel Competition Scheme (PCS) for high channel utilization and energy saving of the battery. In each Competition Cycle (CC), the data packet produced by each SH in a different time slot can join in CC for data packet transmission in parallel for high channel utilization. In CAPL-MAC, each SH is assigned with a different Priority Level Number (PLN) during each CC. Instead of broadcasting, each SH is sending its respective PLN to every SH with the help of nearest SH for saving energy of the battery. Based on highest PLN, each SH is communicating to SN without collision and also it will reduce propagation delay as well as improve timing efficiency. Finally, Quality of Service (QoS) is also improved. We adopt the single layer approach with handshaking protocol for communication. We carried out the simulation utilizing Aqua-Sim Network Simulator2 (NS2). The simulation results illustrate that the proposed CAPL-MAC protocol achieves the earlier stated performance than by the existing protocols such as Competitive Transmission-MAC (CT-MAC) and Channel Aware Aloha (CAA). 
Keywords: Underwater acoustic sensor network, Medium access control protocol, Handshaking protocol, Channel aware aloha, Quality of service.

An online framework for civil unrest prediction using tweet stream based on tweet weight and event diffusion
Md Kamrul Islam, Md Manjur Ahmed, Kamal Zuhairi Zamli, & Salman Mehbub
Twitter is one of most popular internet-based social networking platform to share the feelings, views, and opinions. In recent years, many researchers have utilized the social dynamic property of the posted messages or tweet to predict the civil unrest in advance. However, the existing frameworks fail to describe the low granularity level of tweet and they work in offline mode. Moreover, most of them do not deal with the case where enough tweet information is not available. To overcome these limitations, this article proposes an online framework for analyzing the tweet stream in predicting the future civil unrest events. The framework filters the tweet stream and classifies the tweets using linear SVM classifier. After that, the weight of the tweet is measured and distributed among the extracted locations to update the overall weight in each location in a day in a fully online manner. The weight history is then used to predict the status of civil unrest in a location. The significant contributions of this article are (i) A new keyword dictionary with keyword score to quantify the sentiment for extracting the low granularity level of knowledge (ii) A new diffusion model for extracting the locations of interest and distribute the sentiment among the locations utilizing the concept of information diffusion and location graph to handle the locations with insufficient information (iii) Estimating the probability of civil unrest and finding the stages of unrest in upcoming days. The performance of the proposed framework has been measured and compared with existing logistic regression based predictive framework. The result shows that the proposed framework outperforms the existing framework in terms of F1 score, accuracy, balanced accuracy, false acceptance rate, false rejection rate, Matthews’s correlation coefficient.
KeywordsText classification, information diffusion, sentiment analysis, polynomial regression, connected graph.

Universiti Utara Malaysia Press
 Universiti Utara Malaysia, 06010 UUM Sintok
Kedah Darul Aman, MALAYSIA
Phone: +604-928 4816, Fax : +604-928 4792

All Right Reserved. Copyright © 2010, Universiti Utara Malaysia Press