		FeedBack \| Contact Us \| Links \| Site Map

Welcome to Journal of Theoretical and Applied Information Technology

Submit Paper / Call for Papers

Journal receives papers in continuous flow and we will consider articles from a wide range of Information Technology disciplines encompassing the most basic research to the most innovative technologies. Please submit your papers electronically to our submission system at http://jatit.org/submit_paper.php in an MSWord, Pdf or compatible format so that they may be evaluated for publication in the upcoming issue. This journal uses a blinded review process; please remember to include all your personal identifiable information in the manuscript before submitting it for review, we will edit the necessary information at our side. Submissions to JATIT should be full research / review papers (properly indicated below main title).

Journal of Theoretical and Applied Information Technology
October 2016 | Vol. 92 No.2

Title:	STUDIES ON IMPROVING TEXTURE SEGMENTATION PERFORMANCE USING GENERALIZED GAUSSIAN MIXTURE MODEL INTEGRATING DCT AND LBP
Author:	K. NAVEEN KUMAR, K.SRINIVASA RAO, Y.SRINIVAS, CH. SATYANARAYANA
Abstract:	This paper addresses the performance evaluation of the texture segmentation integrating DCT with LBP. In this method, the whole image is converted in to local binary pattern domain. The LBP image is then divided into different non overlapping blocks. From each block, the DCT coefficients are selected in a zig-zag pattern for each block. Assuming the feature vectors follow a multivariate generalized Gaussian mixture model, the model parameters are estimated using EM algorithm. The initialisation of the model parameters is carried using moment method of estimation and using Hierarchical clustering algorithm. The texture segmentation algorithm is developed under Bayesian frame with component maximum likelihood. The performance of the proposed algorithm is evaluated using performance measures such as GCE, PRI and VOI with randomly selected images from Brodatz database. It is observed that this algorithm outperforms existing texture segmentation algorithms with respect to performance measures.
Keywords:	Texture Segmentation, Multivariate Generalized Gaussian Mixture Model, Performance Measures, Local Binary Patterns, DCT Coefficients.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	SEAT BELT DESIGN BY GRAPHIC APPROACH
Author:	FRIH ABDERRAHIM, CHALH ZAKARIA, MRABTI MOSTAFA, OUAHI MOHAMED
Abstract:	An observer can reconstruct or estimate in real time the current state of a real system using available measures, without prior knowledge of the initial conditions, to have complete and accurate information on system status. In addition, it allows estimating the unmeasurable states of a system, also replacing expensive or difficult to maintain sensors. For example the slip sensor.... For this aim, we will conceive full-order Luenberger observer on a linear system using a graphical approach to model it and that will be easy to handle thereafter, the causal properties and structural analysis of the bond graph tool are implemented for the synthesis of the observer. As a result, the calculation of the gains of the Luenberger observer is based on graphical procedures. Also to give a comparison between the measured values and estimate to demonstrate the importance and efficiency of the proposed approach. Simulation tests on a design of the seat belt field shows the performance of the observer Luenberger linear graph developed in order to measure and estimate the state unmeasurable.
Keywords:	Bond Graph, Luenberger Observer, Belt Field, Performance, Estimate.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	AN EFFICIENT WEB USAGE MINING ALGORITHM BASED ON LOG FILE DATA
Author:	TAWFIQ A. AL-ASDI, AHMED J. OBAID
Abstract:	Information on Internet and specially on website environment is increasing rapidly day by day and become very huge, this information play an important role for discovering various knowledge in the Web. Web Usage Mining one of the Web Mining algorithm categories that concern with discover and analysis useful information regard to link prediction, users' navigation, customers' behavior, site reorganization, web personalization and frequent access patterns from large web data that logs by Web server side and stored in standard text log file format called log file or Web usage data, this data can also be collected from an organization's database such as NASA. Web Usage Mining is a process of applying Data mining techniques and application to analyze and discover interesting knowledge from the Web. There are several existing research works on log file mining, some concern with web site structure, traversal pattern mining, association rule mining, Web page classification, and general statistics such as amount of time spent on a page. In this paper we will focus on mining the different segments content of Web log data entries in order to discover the hidden information and interesting browsing contents from it, then applying clustering algorithm to find similar groups of Web sites that have common browsing contents.
Keywords:	Web Mining, Web Usage Mining, Log File Analysis, Clustering, K-means, System Monitoring.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	PROVIDING A SOLUTION TO IMPROVE PRE-COPY METHOD FOR MIGRATING VIRTUAL MACHINES IN CLOUD INFRASTRUCTURE
Author:	PARVIN AHMADI DOVAL AMIRI, SHAYAN ZAMANI RAD, FARAMARZ SAFI ISFAHANI
Abstract:	Cloud computing can be defined as a new computing model which suggests solutions for providing information technology services analogous to utility services such as power electricity, telephone. Thanks to the virtualization technology, most of Cloud datacenters use this technology for performance improvement. This technology transforms a physical server to several virtual machines. With regard to this property, virtual machines can be transferred from one place to another place that is called migration. Pre-Copy is one of the main methods in migration techniques. An important problem in this method is that when the memory pages are changed (dirty pages) faster than sending pages to a destination, the process of migration will take a long time that is the case in transferring virtual machines. In this paper, a solution is proposed that sets up the speed of virtual machines in terms of CPU frequency and makes a balance between senders and receivers in pre-copy algorithm. The results show that the proposed method has better performance compared to the pre-copy method and decreases the total migration time around 25% and the total amount of transferred data around 47%.
Keywords:	Cloud Computing, Migration, Virtualization, Virtual Machine
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	DISCRETE RISK MODELS OF THE PROCESS OF VIRAL EPIDEMICS DEVELOPMENT IN HOMOGENOUS INFORMATION AND TELECOMMUNICATION NETWORKS
Author:	ELENA NIKOLAEVNA PONOMARENKO, VERA NIKOLAYEVNA KOSTROVA, RUSLAN KALANDAROVICH BABADZHANOV, YURIY NIKOLAEVICH GUZEV, VLADIMIR SERGEYEVICH ZARUBIN
Abstract:	The object of this research is homogeneous networks, i.e., networks with small fluctuations of the vertices degree. In other words, in such homogeneous structures assume that k≅〈k〉, where parenthesis mean averaging over the degree of distribution. The structures also have a very specific application in practice. Typically, such networks require stringent communication within any corporation. This network organization is particularly relevant in information and telecommunication networks (ITN) of critically important objects. In this paper, the models such as SI, SIS, SEIS, SIR, SEIR, MSEIR are distinguished, when a malware attack is performed by network virus, which uses vulnerability in the work of network services of the operating system for its spread. The model synthesis is performed under the conditions that ITN contacts can be represented by a complete graph, and the epidemic occurs in a closed ITN. The features of the process of malware program transmission from one computer to another, as well as the internal features of malware program execution on the computer, are ignored. For each model, the analytical expressions of epidemic resistance were obtained. The prospects of using the proposed models in the process of network warfare were distinguished.
Keywords:	Epidemic, Risk, Epidemic Resistance, Mathematical Expectation, Mode, Standard Deviation (SD), Information And Telecommunication Network (ITN).
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	ON THE IDENTIFICATION OF THE STRUCTURAL PATTERN OF TERMS OCCURRENCE IN A DOCUMENT USING BAYESIAN NETWORK
Author:	SOEHARDJOEPRI, NUR IRIAWAN, BRODJOL SUTIJO SU, IRHAMAH
Abstract:	The pattern of text documents is strongly influenced by the advent of the first term in composing term structure of each sentence. When two documents have the same pattern, then the second and the following terms tend to be same. This paper would create a special tool for detecting the similarity of structural pattern of two text documents. Latent Semantic Analysis(LSA) couples with Bayesian Network (BN) are employed as the main engine to build the algorithms. The work of these approaches is demonstrated to detect the similarities of the appearance of the term in the sentence in any text documents.
Keywords:	Text Pattern Document, Term, Latent Semantic Analysis, Bayesian Network.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	A NOVEL TOP-K INFREQUENT MINING TECHNIQUE ON COMPLEX DISTRIBUTED MARKET DATABASES
Author:	SUJATHA KAMEPALLI, RAJA SEKHARA RAO KURRA, SUNDARA KRISHNA .Y.K
Abstract:	Infrequent association rule mining is one of the essential tasks in data mining research to find rare items on complex data set. Also, most of the traditional models focus on finding negative association rules based on different association measures. However, finding relational infrequent patterns from the large number of candidate sets is still an open problem in the distributed market analysis. Traditional infrequent mining models are mainly depending on quantitative attributes, limited data size and Boolean datasets. In any distributed environment, as the size and complexity of the market data increases, it is difficult to find the sparsity issue from the positive association rules. In this proposed approach a novel infrequent association mining algorithm was implemented to find the topmost relational infrequent patterns from the complex market dataset. Experimental outcomes prove that the proposed model extracts high quality, infrequent patterns compared to conventional infrequent rule mining techniques.
Keywords:	Complex Data, Infrequent Association Rules, Data Sparsity, Quantitative Association Rules, Rank Correlation.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	COURSE SCHEDULING USING MULTI-AGENT EXPLORATION METHOD
Author:	JAMALUDIN HAKIM, RETANTYO WARDOYO, SRI HARTATI, AHMAD ASHARI
Abstract:	The issue of course scheduling is a frequent topic, involving the use of new techniques with promising results. This research used exploration method on the Schedule Media with the agents providing a flexible schedule based on the resulted combination and a stable process. The constraints include lecturers’ teaching preferences and students’ different choices of courses in the Semester Credit System; which would be the problem being addressed in this study. Using multi-agents, the exploration method would be able to process lecturers’ constraints and varying courses chosen by students from varying semesters. Agents’ responses to requests during the scheduling process would determine the success of the scheduling.
Keywords:	Scheduling, Exploration Method, Constraints, Multi-Agent, Schedule Media
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	NEURAL NETWORK OPTIMIZATION USING SHUFFLEDFROG ALGORITHM FOR SOFTWARE DEFECT PREDICTION
Author:	REDDI. KIRAN KUMAR, S.V.ACHUTA RAO
Abstract:	Software Defect Prediction (SDP) focuses on the detection of system modules such as files, methods, classes, components and so on which could potentially consist of a great amount of errors. SDP models refer to those that attempt to anticipate possible defects through test data. A relation is present among software metrics and the error disposition of the software. To resolve issues of classification, for the past many years, Neural Networks (NN) have been in use. The efficacy of such networks rely on the pattern of hidden layers as well as in the computation of the weights which link various nodes. Structural optimization is performed in order to increase the quality of the network frameworks, in two separate cases: The first is the typically utilized approximation error for the present data, and the second is the capacity of the network to absorb various issues of a general class of issues in a rapid manner along with excellent precision. The notion of Back Propagation (BP) is quite elementary; the result of neural networks is tested against the desired outcome. Genetic algorithms (GA) are a type of search algorithms built, based on the idea of natural evolution. A neural network using Shuffled Frog Algorithm for improving SDP is proposed.
Keywords:	Software Defect Prediction (SDP), Neural Network (NN), Back Propagation (BP), Genetic Algorithm (GA) and Shuffled Frog Leaping Algorithm (SFLA)
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	SYNTHETIC RANGE IMAGE SIMULATION OF TERRESTRIAL LIDAR SCANNER
Author:	THINAL RAJ, FAZIDA HANIM HASHIM
Abstract:	In recent years, the usage of Light Detection and Ranging (LiDAR) scanners has become predominant in the fields of robotics, navigation and remote sensing for numerous applications. Due to the high selling cost of LiDAR scanners, the research communities have developed LiDAR simulators mostly for Airborne Laser Scanning (ALS) application and some minimal works can be found for Terrestrial Laser Scanning (TLS) application. While the LiDAR simulators are becoming more sophisticated, there are still some applications area where the use of these simulators are not convenient. This paper describes an alternative and simple approach for developing a simulator for terrestrial LiDAR scanner application. The aim of this research is to model a synthetic 2D LiDAR scanner to provide range data based on object and sensor parameters input provided by the user. The object models are derived explicitly from geometrical relationships between the object and ray from scanner, thus avoiding the use of vectors. This approach directly raster the range image in contrast to ray tracing. The GUI-based simulator is developed in Processing language, thus it has an open source and hardware independent environment, making it ideal for research applications. The data obtained from this simulator can be used as ground truth for analyzing the characteristics of LiDAR scanners or can even be used for localization simulations.
Keywords:	Terrestrial LiDAR, 2D Laser Scanner, Simulator, Range Image, Synthetic, Range Data
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	HAM: A HYBRID ALGORITHM PREDICTION BASED MODEL FOR ELECTRICITY DEMAND
Author:	WAHAB MUSA, SALMAWATY TANSA
Abstract:	Based on the rapid development of digital signal processing technology and computers, computational intelligence (CI) becomes an object of study fields of fundamental and applied research of interest to researchers recently. Research that exploits a number of further processing techniques are the subject of CI information technology, including artificial neural networks, genetic algorithms, fuzzy logic, evolutionary computation, or a combination of these techniques, known as hybrid technology. The purpose of this study is to develop a prediction model based CI to improve the accuracy of prediction of the demand for electricity. Method combination of several algorithms in search of optimum value will be developed to overcome the premature convergence on the model predictions. The data will be used to measure the performance of the hybrid model is data electrical energy needs of Indonesia. Average prediction errors will become a reference in selecting the right model for the planning of the electrical energy needs of the next few years in Indonesia. The results showed: 1) performance computational intelligence-based prediction models that utilize the capabilities of the hybrid algorithm is superior to the single algorithm-based predictive models. 2) The accuracy of the prediction model based hybrid algorithm (HAM) can reach 97%, exceeding the level of the previous model's accuracy.
Keywords:	Computational Intelligence, Hybrid Algorithm, Prediction Model.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	UNDERSTANDING NETWORK CONGESTION EFFECTS ON PERFORMANCE - ARTICLES REVIEW
Author:	MOHAMED NJ., SHAHRIN SAHIB, NANNA SURYANA, BURAIRAH HUSSIN
Abstract:	Networking communications have become popular worldwide in human daily services. Network Congestion (NC) happens whenever because nodes and links are overloaded. Such situations affect the network expected performance and its services quality. Congestion /NC occurs as a results of its subnets’ links overload, which gradually (overtime) affects the network performance with an increase of transmission delay, a slowdown of throughput as generally perceived by network’s users. NC is considerable as the basis problem in network performance quality acceptance; and most of its existing problem solutions are expected still playing a great role in the future networks model, which will be running mostly too many multimedia applications. However, various researches over past years have initiated the study on the causes leading to congestion and, different lessons can be learnt from NC situations analysis to understand its relationship with the future network’s performance. This paper presents an analytical review of NC occurrence causes and the fundamentals of the existing control solutions/frameworks as available and studied from some former and recent networks publications. A particular attention has been paid throughout this study to found out how NC may still affect the future networks performance (i.e. QoS in the world of multimedia networks). And the coverage/content of this paper is expected to serve as a quick access to the knowledge essentials for researchers on related subject as stated in this paper topic.
Keywords:	Communication, congestion, performance, service quality, QoS, network, networking, WLAN/LAN, choke point, congestion, edge router, flow, traffic, data rate, wireless.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	IMPLEMENTATION OF SENSORLESS CONTROL OF AN INDUCTION MOTOR ON FPGA USING XILINX SYSTEM GENERATOR
Author:	SGHAIER NARJESS, TRABELSI RAMZI, MIMOUNI MED FAOUZI
Abstract:	In this paper, we will presented a deterministic observation approach , nonlinear applied to the Induction Motor: this is a sliding mode observer. Indeed, this paper will serve to emphasize the importance of the order without sensor to increase the profitability of our machine. Our sliding mode observer will be applied to the field oriented vector control then to the sliding mode control. The contribution of this paper is the design of sensorless control using XSG order to be implanted on the FPGA.
Keywords:	Induction motor, FPGA, Sensorless Control, Sliding mode observer, XSG
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	AN ARTIFICIAL BEE COLONY BASED OPTIMIZATION FOR IMAGE SEGMENTATION IN MARKOVIAN FRAMEWORK: APPLICATION TO NON DESTRUCTIVE TESTING IMAGES
Author:	MOHAMED BOU-IMAJJANE, MOHAMED SBIHI
Abstract:	Image segmentation is a fundamental task in image analysis process. In this paper, we propose a segmentation model using MRF (Markov Random Fields) and a global optimization method based on ABC (Artificial Bee Colony) algorithm. As a Markovian algorithm, ICM (Iterated Conditional Modes) is a segmentation method which takes into account the neighbouring labels of the pixel in calculating the energy function that need to be minimized to obtain the best segmentation. Though, in some cases, this segmentation method may converge to the first encountered minimum during the minimization process. To face out this situation, ABC is so used to improve the energy function optimization process since it gives robust results especially in discrete multivariable optimization problems. The contribution of this work is to propose MRF-ABC algorithm that consists of introducing ABC to optimize Markovian energy function at the convergence point obtained using ICM to adjust the belongingness of pixels in order to improve image segmentation quality for X-ray Non Destructive Testing images.
Keywords:	Markov Random Fields, Artificial Bee Colony, Image Segmentation, Potts Energy Function, Non Destructive Testing
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	ENHANCING THE AD-HOC MESH NETWORKS INFRASTRUCTURE IN RURAL AREAS: ADAPTIVE APPROACH
Author:	IBRAHIM OBEIDAT, ALI ABU ABID, HUTAF NATOUREAH
Abstract:	Rural areas all over the world continue to be poorly covered and are not considered as an achievable business case by telecommunication operators. This is due to high implementation costs compared to the profit. The growth in telecommunications, as well as new mobile technology has expanded the gap between rural and urban areas in networks’ infrastructure. To solve this issue, this research uses ad-hoc mesh networks, where each node has a transmission range of one kilometer, which means it needs a third party to play the role of relay if the distance is more than one kilometer. Findings of this research suggest that if the area coverage is less than or equal to 5%, the number of relay nodes reduces by 50%, whereas if the reduction in relay nodes is 8.5% when the area coverage is more than 5%, then the second assumption is more precise and effective.
Keywords:	Ad-Hoc Mesh Networks, Connectedness, Terranet, Adjacency Matrix, Adaptive Threshold
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	FEATURE IDENTIFICATION OF AN INTERACTIVE MULTI-USER MEETING TABLETOP BASED ON INTUITIVE GESTURE RECOGNITION
Author:	HALEEMA SADIA and ABDULLAH MOHD ZIN
Abstract:	Multi-touch technology has shown a rapid rise in popularity over the last few years, being implemented in many devices from interactive walls to interactive tables and from mobile phones to desktop monitors. It has provided users with an extremely intuitive means of interaction with electronic devices through gesture based self-sensing control. The advances in touch technology have led to an increased presence of multi-touch interfaces in consumer products in recent years. However, very little research has been done in developing interactive multi touch multi user project management system. The interactive multi-touch multi-user project management system based on tabletop surface helps users to execute their important meetings in a more productive and less intrusive manner. It also facilitates new ways to foster collaborative creation, permitting several users to work simultaneously on a single screen. This paper describes some of the features that are required for such an application. The feature identification process is carried out through comparative analysis and evaluation of scenario analysis. This paper also discusses some of the issues and challenges of multi-touch technology.
Keywords:	Multi-user, Multi-touch Tabletop, Project management, Project planning
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	XML INTEGRITY CONSTRAINTS, WHAT’S NEXT?
Author:	MOHAMMED HAKAWATI, PUTEH SAAD, NASEER SABRI, YASMIN YACOB, R. B. AHMAD, M. S. SALIM
Abstract:	Without any doubt, XML data model considered the most dominant document type over the web with more than 60% of the total; nevertheless, their quality is not as expected. Data cleaning is equipped to overcome database’s quality issues. Integrity Constraint is a very important criterion for keeping data in a consistent manner, almost all previous XML dependencies are introduced to improve the schema and normalization, with a small effort toward improving data instance. This paper summarizes the most important XML integrity constraint notations and data cleaning approaches. In addition, to highlight the shortcoming of these constraints and proved it is limitation for increasing data quality. Finally, introduce the next generation of conditional integrity constraints, which will be held mainly for data cleaning issues.
Keywords:	XML, Data Quality, Data Cleaning, Integrity Constraints.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	KLANG VALLY RAINFALL FORECASTING MODEL USING TIME SERIES DATA MINING TECHNIQUE
Author:	ZULAIHA ALI OTHMAN, NORAINI ISMAIL, ABDUL RAZAK HAMDAN, MAHMOUD AHMED SAMMOUR
Abstract:	Rainfall has influence the social and economic activities in particular area such as agriculture, industry and domestic needs. Therefore, having an accurate rainfall forecasting becomes demanding. Various statistical and data mining techniques are used to obtain the accurate prediction of rainfall. Time series data mining is a well-known used for forecast time series data. Therefore, the objective of this study is to develop a distribution of rainfall pattern forecasting model based on symbolic data representation using Piecewise Aggregate Approximation (PAA) and Symbolic Aggregate approXimation (SAX). The rainfall dataset were collected from three rain gauge station in Langat area within 31 years. The development of the model consists of three phases: data collection, data pre-processing, and model development. During data pre-processing phase, the data were transform into an appropriate representation using dimensional reduction technique known as Piecewise Aggregate Approximation (PAA). Then the transformed data were discretized using Symbolic Aggregate approXimation (SAX). Furthermore, clustering technique was used to determine the label of class pattern during preparing unsupervised training data. Three type of pattern are identified which is dry, normal and wet using three clustering techniques: Agglomotive Hierarchical Clustering, K-Means Partitional Clustering and Self-Organising Map. As a result, the best model has be able to forecast better for the next 3 and 5 years using rule induction classification techniques.
Keywords:	Time Series Data Mining, Clustering, Classification, Time Series Symbolic Representation, Rainfall Forecasting
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	IMPROVED FAULT DETECTION IN WATER DESALINATION SYSTEMS USING MACHINE LEARNING TECHNIQUES
Author:	MORCHED DERBALI, ANAS FATTOUH, HOUSSEM JERBI, MOHAMED NACEUR ABDELKRIM
Abstract:	In this paper, the authors have attempted to study the fault detection using the machine learning technique for the water Membrane Distillation Systems (MDS). Initially, an actual system with the MDS, applying nanotechnology was developed which was based on actual measurements. Then, the errors occurring between the outputs of the model (additionally, these outputs serve as MDS inputs) and system outputs were classified for identifying the system faults. This type of classification was carried out by using different approaches and the classification results were further compared. It was noted that the classification accuracy obtained by using the decision trees was the best as compared to the other learning techniques like K-Nearest Neighbours, Neural Networks, and the Support Vector Machines (SVM).
Keywords:	Learning Techniques, Water Membrane Distillation System, Desalination Systems, Fault Detection, Detection Accuracy.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	WAVELET DECOMPOSITION LEVELS ANALYSIS FOR INDONESIA TRADITIONAL BATIK CLASSIFICATION
Author:	FIKRI BUDIMAN, ADANG SUHENDRA, DEWI AGUSHINTA, AVINANTA TARIGAN
Abstract:	Indonesia traditional batik is a non-material cultural heritage which has patterns that basically divided into batik ‘keraton/pedalaman’ and ‘pesisir’. An appropriate content-based image recognition method is needed to recognize the pattern of Indonesa traditional batik in a large image database. The results of this research can be used to recognize batik ‘keraton’ and ‘pesisir’ based on feature of wavelet energy and standard deviation, with 122 images as training dataset and 120 images as test datasets. Classification using binary non-linear support vector method, with feature extraction of discrete wavelete Transform (DWT) which was tested for wavelet types of Daubechies 1 - 5 with decomposition first level, and Daubechies 2 with decomposition 1st to 5th level. The Best result is obtained by Daubechies 2 decomposition thirth level with an accuracy of 96.7%. The result is better than the previous researches with the same datasets and classification method. The previous researches conducted using feature extraction method with fractal feature obtained an accuracy at 91.6%, and that which used GLCM with 20 parameters obtained an accuracy at 80%.
Keywords:	Image Retrieval, Traditional Batik, Decomposition Wavelet, Feature Extraction,Classification.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	K-NEAREST NEIGHBOR BASED DBSCAN CLUSTERING ALGORITHM FOR IMAGE SEGMENTATION
Author:	SURESH KURUMALLA, P SRINIVASA RAO
Abstract:	Clustering is a primary and vital part in data mining. Density based clustering approach is one of the important technique in data mining. The groups that are designed depending on the density are flexible to understand and do not restrict itself to the outlines of clusters. DBSCAN Algorithm is one of the density grounded clustering approach which is employed in this paper. The author addressed two drawbacks of DBSCAN algorithm i.e. determination of Epsilon value and Minimum number of points and further proposed a novel efficient DBSCAN algorithm as to overcome this drawback. This proposed approach is introduced mainly for the applications on images as to segment the images very efficiently depending on the clustering algorithm. The experimental results of the suggested approached showed that the noise is highly reduced from the image and segmentations of the images are also improved better compared to the existing image segmentation approaches.
Keywords:	Data Mining, Clustering, Density Based Clustering, DBSCAN, K-Nearest Neighbor, Image Segmentation.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text

Title:	CLASSIFICATION OF VOIP AND NON-VOIP TRAFFIC USING MACHINE LEARNING APPROACHES
Author:	GHAZI AL-NAYMAT, MOUHAMMD AL-KASASSBEH, NOSAIBA ABU-SAMHADANH, SHERIF SAKR
Abstract:	Enhancing network services and security can be achieved by performing network traffic classification identifying applications, which is one of the primary components of network operations and management. The traditional transport-layer and port-based classification approaches have some limitations in achieving accurate identification. In this paper, a real test bed is used to collect first-hand traffic dataset from five different VoIP and Non-VoIP applications that are used by majority of Internet community, namely Skype, YouTube, Yahoo Messenger, GTalk and PayPal. The collected data encompasses new features that have never been used before. In addition, a classification step is performed using off-the-shelf machine learning techniques, specifically Random Forest J48, meta.AdaBoost (J48) and MultiLayer Perceptron to classify the traffic. Our experimental results show that using the new features can dramatically improve the true positive ratio by up to 98% and this is significant outcome towards providing accurate traffic classification.
Keywords:	Traffic classification, Application identification, Machine Learning, VOIP and Non-VOIP Application, CAPTCHA
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2016 -- Vol. 92. No. 2 -- 2016
Full Text