Sunday, December 17, 2017
   
Text Size
Login

Journal of Information and Communication Technology (JICT) Vol. 14 2015

COST-SENSITIVE STRUCTURED PERCEPTRON INCORPORATING CATEGORY HIERARCHY FOR NAMED ENTITY RECOGNITION
¹Shohei Higashiyama, ²Blondel Mathieu, ³Kazuhiro Seki & Kuniaki Uehara
¹Graduate School of System Informatics
Kobe University, Japan
 
²NTT Communication Science Laboratories, Kobe University, Japan
 
³Faculty of Intelligence and Informatics, Konan University, Japan
higashiyama@ai.cs.kobe-u.ac.jp; mathieu@blondel.org; seki@konan-u.ac.jp;uehara@kobe-u.ac.jp
 
ABSTRACT FULL TEXT
Named Entity Recognition (NER) is a fundamental natural language processing task for the identifi cation and classifi cation of expressions into predefi ned categories, such as person and organization. Existing NER systems usually target about 10 categories and do not incorporate analysis of category relations. However, categories often belong naturally to some predefi ned hierarchy. In such cases, the distance between categories in the hierarchy becomes a rich source of information that can be exploited. This is intuitively useful particularly when the categories are numerous. On that account, this paper proposes an NER approach that can leverage category hierarchy information by introducing, in the structured perceptron framework, a cost function more strongly penalizing category predictions that are more distant from the correct category in the hierarchy. Experimental results on the GENIA biomedical text corpus indicate the effectiveness of the proposed approach as compared with the case where no cost function is utilized. In addition, the proposed approach demonstrates the superior performance over a representative work using multi-class support vector machines on the same corpus. A possible direction to further improve the proposed approach is to investigate more elaborate cost functions than a simple additive cost adopted in this work.
 
Keywords: Named entity recognition, category hierarchy, cost-sensitive learning, biomedical text mining.
 

 
HYBRID NSGA-II OPTIMIZATION FOR IMPROVING THE THREE-TERM BP NETWORK FOR MULTICLASS CLASSIFICATION PROBLEMS
¹,²Ashraf Osman Ibrahim, ²Siti Mariyam Shamsuddin & ³Sultan Noman Qasem
¹Faculty of Computer and Technology, Alzaiem Alazhari University,Khartoum North, Sudan
 
²UTM Big Data Centre, Universiti Teknologi Malaysia, Malaysia
 
³College of Computer and Information Sciences, Al Imam Mohammad Ibn
Saud Islamic University (IMSIU), Riyadh, Saudi Arabia
ashrafosman2@gmail.com; mariyam@utm.my; sultann.noman@ccis.imamu.edu.sa
 
ABSTRACT FULL TEXT
Recently, hybrid algorithms have received considerable attention from a number of researchers. This paper presents a hybrid of the multiobjective evolutionary algorithm to gain a better accuracy of the fi nal solutions. The aim of using the hybrid algorithm is to improve the multiobjective evolutionary algorithm performance in terms of the enhancement of all the individuals in the population and increase the quality of the Pareto optimal solutions. The multiobjective evolutionary algorithm used in this study is a nondominated sorting genetic algorithm-II (NSGA-II) together with its hybrid, the backpropagation algorithm (BP), which is used as a local search algorithm to optimize the accuracy and complexity of the three-term backpropagation (TBP) network. The outcome positively demonstrates that the hybrid algorithm is able to improve the classification performance with a smaller number of hidden nodes and is effective in multiclass classifi cation problems. Furthermore, the results indicate that the proposed hybrid method is a potentially useful classifi er for enhancing the classification process ability when compared with the multiobjective genetic algorithm based on the TBP network (MOGATBP) and certain other methods found in the literature.
 
Keywords: Artificial Neural Network, hybridization technique, genetic algorithm, NSGA-II, multiobjective optimization.
 

 
A HYBRID SIMULATION APPROACH IN DEVELOPING A RISK QUANTIFICATION MODEL FOR COAL PROCUREMENT IN POWER GENERATION
¹Jafni Azhan Ibrahim, ²Hasimah Sapiri & ³Razman Mat Tahar
¹School of Technology Management and Logistic,
Universiti Utara Malaysia, Malaysia
 
²School of Quantitative Sciences, Universiti Utara Malaysia, Malaysia
 
³Faculty of Technology Management, Universiti Malaysia Pahang, Malaysia
jafni@uum.edu.my; hsimah@uum.edu.my; razman779@ump.edu.my
 
ABSTRACT FULL TEXT
All energy systems provide some level of security to its consumers. However, the right or optimum level of security is very diffi cult to assess. In order to make a comparison between the cost of providing energy security and level of security, the quantifi ed risks have to be in the common accounting platform of cost. Then, an optimum level of security and cost can be estimated using appropriate methods. Since no such attempts have been done to compare the risks and the cost in the same platform, i.e. monetary unit, it is unpractical to determine the optimum point between the risks and cost of providing security in any energy systems. The objective of this paper is to present a new hybrid simulation model in risk analysis which computes the total exposure of coal procurement in power generation through the summation of quantifi ed supply shortage risk in monetary terms and the cost of coal procurement. The hybrid simulation model is made up of two main components: 1) Dynamic Risk Calculation Program (DR-P) which was developed in System Dynamics platform for capturing the effect of dynamics behavior of price toward coal procurement risk, and 2) Non-delivery Probability Table Program (NdPT-P) which was developed in Matlab platform for computing all possible shortage level and the probability of shortage in selected coal procurement portfolio. The result from this paper has shown that the risks of coal procurement were increased as the cost of coal procurement were decreased and vice versa. However, the summation of risk and cost which give the total exposure of coal procurement has provided more accurate information for selecting the best coal procurement portfolio option.
 
Keywords: Coal procurement, power generation, risk quantifi cation, dynamics simulation, system dynamics.
 

 
IMPROVED SPEAKER-INDEPENDENT EMOTION RECOGNITION FROM SPEECH USING TWO-STAGE FEATURE REDUCTION
¹Hasrul Mohd Nazid, ²Hariharan Muthusamy, ³Vikneswaran Vijean
¹, ², & ³ School of Mechatronic Engineering,
Universiti Malaysia Perlis, Malaysia
Universiti Kuala Lumpur Malaysian Spanish Institute,
 
Sazali Yaacob
Kulim Hi-Tech Park, Malaysia
kids.hasrul@gmail.com;wavelet.hari@gmail.com;viky.86max@gmail.com;sazali22@yahoo.com
 
ABSTRACT FULL TEXT
In the recent years, researchers are focusing to improve the accuracy of speech emotion recognition. Generally, high emotion recognition accuracies were obtained for two-class emotion recognition, but multi-class emotion recognition is still a challenging task . The main aim of this work is to propose a two-stage feature reduction using Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for improving the accuracy of the speech emotion recognition (ER) system. Short-term speech features were extracted from the emotional speech signals. Experiments were carried out using four different supervised classifi ers with two different emotional speech databases. From the experimental results, it can be inferred that the proposed method provides better accuracies of 87.48% for speaker dependent (SD) and gender dependent (GD) ER experiment, 85.15% for speaker independent (SI) ER experiment, and 87.09% for gender independent (GI) experiment.
 
Keywords: Emotional speech, cepstral features, feature reduction, emotion recognition
 

.
PROTECTING HOME AGENT CLIENT FROM IPv6 ROUTING HEADER VULNERABILITY IN MIXED IP NETWORKS
¹Bassam Naji Al-Tamimi, ²Mohamed Shenify & ³Rahmat Budiarto
¹Taibah University, Almadinah Almonawarah, Saudi Arabia
 
²&³Albaha Uniersity, Albaha, Saudi Arabia
btamimi@taibahu.edu.sa; maalshenify@bu.edu.sa; rahmat@bu.edu.sa
 
ABSTRACT FULL TEXT
Mixed IPv4/IPv6 networks will continue to use mobility support over tunneling mechanisms for a long period of time until the establishment of IPv6 end-to-end connectivity. Encapsulating IPv6 traffi c within IPv4 increases the level of hiding internal contents. Thus, mobility in mixed IPv4/IPv6 networks introduces new security vulnerabilities. One of the most critical vulnerabilities associated with the IPv6 protocol is the routing header that potentially may be used by attackers to bypass the network security devices. This paper proposes an algorithm (V6HAPA) for protecting home agent clients from the routing header vulnerability, considering that the home agents reside behind an IPv4 Network Address Translation (NAT) router. The experimental results show that the V6HAPA provides enough confidence to protect the home agent clients from attackers.
 
Keywords: Mobile IP, IPv4/IPv6 coexistence, IPv6 security, IPv6 routing header.
 

 
IRRELEVANT FEATURE AND RULE REMOVAL FOR STRUCTURAL ASSOCIATIVE CLASSIFICATION
¹Izwan Nizal Mohd Shaharanee & ²Jastini Mohd Jamil
¹&²School of Quantitative Sciences, Universiti Utara Malaysia, Malaysia
nizal@uum.edu.my; jastini@uum.edu.my
 
ABSTRACT FULL TEXT
In the classification task, the presence of irrelevant features can significantly degrade the performance of classification algorithms, in terms of additional processing time, more complex models and the likelihood that the models have poor generalization power due to the over fitting problem. Practical applications of association rule mining often suffer from overwhelming number of rules that are generated, many of which are not interesting or not useful for the application in question. Removing rules comprised of irrelevant features can signifi cantly improve the overall performance. In this paper, we explore and compare the use of a feature selection measure to filter out unnecessary and irrelevant features/attributes prior to association rules generation. The experiments are performed using a number of real-world datasets that represent diverse characteristics of data items. Empirical results confirm that by utilizing feature subset selection prior to association rule generation, a large number of rules with irrelevant features can be eliminated. More importantly, the results reveal that removing rules that hold irrelevant features improve the accuracy rate and capability to retain the rule coverage rate of structural associative association.
 
Keywords: Features selection, rules removal, frequent item set mining.
 

 
SOFT BIOMETRICS: GENDER RECOGNITION FROM UNCONSTRAINED FACE IMAGES USING LOCAL FEATURE DESCRIPTOR
Olasimbo Ayodeji Arigbabu, Sharifah Mumtazah Syed Ahmad, Wan Azizun Wan Adnan & Saif Mahmood
Department of Computer and Communication Systems,
Universiti Putra Malaysia, Malaysia
 
Salman Yussof
Department of Systems and Networking,
Universiti Tenaga National, Malaysia
oa.arigbabu@gmail.com; s_mumtazah@upm.edu.my; wawa@upm.edu.my; salman@uniten.edu.my; saiffd2000@gmail.com
 
ABSTRACT FULL TEXT
Gender recognition from unconstrained face images is a challenging task due to the high degree of misalignment, pose, expression, and illumination variation. In previous works, the recognition of gender from unconstrained face images is approached by utilizing image alignment, exploiting multiple samples per individual to improve the learning ability of the classifier, or learning gender based on prior knowledge about pose and demographic distributions of the dataset. However, image alignment increases the complexity and time of computation, while the use of multiple samples or having prior knowledge about data distribution is unrealistic in practical applications. This paper presents an approach for gender recognition from unconstrained face images. Our technique exploits the robustness of local feature descriptor to photometric variations to extract the shape description of the 2D face image using a single sample image per individual. The results obtained from experiments on Labeled Faces in the Wild (LFW) dataset describe the effectiveness of the proposed method. The essence of this study is to investigate the most suitable functions and parameter settings for recognizing gender from unconstrained face images.
 
Keywords: Gender recognition, unconstrained face images, soft biometric traits, local feature descriptor, shape feature extraction.
 

 
AN ANALYSIS OF EDUCATIONAL GAMES DESIGN FRAMEWORKS FROM SOFTWARE ENGINEERING PERSPECTIVE
¹Mifrah Ahmad, ²Lukman AB. Rahim & ³Noreen Izza Arshad
¹, ² & ³Computer and Information Sciences Department,
Universiti Teknologi Petronas, Malaysia
mifrah09@gmail.com; lukmanrahim@petronas.com.my; noreenizza@petronas.com.my
 
ABSTRACT FULL TEXT
Game-based learning has dominantly embedded itself into a tool of education in the 21st century. In developing educational games, many researchers have proposed frameworks to defi ne elements of an educational game. This paper presents a survey of the different frameworks for educational games and analyzes these frameworks against several criteria for effective video games, well-designed games and key elements of educational games. The authors will also look at the frameworks support towards learning theories. In addition, the analysis continues in the context of software engineering practices to develop effective educational games.
 
Keywords: Game design frameworks, educational game components, software engineering, game design framework.
 

 
ASSESSMENT OF STUDENTS’ COGNITIVE–AFFECTIVE STATES IN LEARNING WITHIN A COMPUTER-BASED ENVIRONMENT: EFFECTS ON PERFORMANCE
¹Ruili Wang, ²Hokyoung Ryu & ³Norliza Katuk
¹School of Engineering and Advanced Technology (SEAT), Massey
University Auckland, New Zealand
 
²Department of Industrial Engineering, Hanyang University Seoul, Korea
 
³School of Computing, Universiti Utara Malaysia
r.wang@massey.ac.nz; hokyoung.ryu@gmail.com; k.norliza@uum.edu.my
 
ABSTRACT FULL TEXT
Students’ cognitive-affective states are human-elements that are crucial in the design of computer-based learning (CBL) systems. This paper presents an investigation of students’ cognitiveaffective states (i.e., engaged concentration, anxiety, and boredom) when they learn a particular course within CBL systems. The results of past studies by other researchers suggested that certain cognitive-affective states; particularly boredom and anxiety could negatively infl uence learning in a computer-based environment. This paper investigates the types of cognitive-affective state that students experience when they learn through a specifi c instance of CBL (i.e., a content sequencing system). Further, research was carried to understand whether the cognitive-affective states would infl uence students’ performance within the environment. A one-way between-subject-design experiment was conducted utilizing four instruments (i) CBL systems known as IT-Tutor for learning computer network, (ii) a pre-test, (iii) a post-test, and (iv) self-report inventory to capture the students’ cognitive-affective states. A cluster analysis and discriminant function analysis were employed to identify and classify the students’ cognitiveaffective states. Students were classifi ed according to their prior knowledge to element the effects of it on performance. Then, non-parametric statistical tests were conducted on different pairs of cluster of the cognitive-affective states and prior knowledge to determine differences on students’ performance. The results of this study suggested that all the three cognitive-affective states were experienced by the students. The cognitive-affective states were found to have positive effects on the students’ performance. This study revealed that disengaged cognitive-affective states, particularly boredom can improve learning performance for lowprior knowledge students.
 
Keywords: Computer-based learning, cognitive-affective states, learning engagement, learning experience.

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