Combination of Adaptive Resonance Theory 2 and RFM Model for Customer Segmentation In Retail Company.

Sr. No

Topic

1.

Scope of the Journal

2.

The Model

3.

The Advisory and Editorial Board

4.

Papers

First Published in the United States of America. Copyright © 2012

Foundation of Computer Science Inc.

International Journal of Computer Applications (IJCA) creates a place for publication of papers which covers the frontier issues in Computer Science and Engineering and their applications which will define new wave of breakthroughs. The journal is an initiative to identify the efforts of the scientific community worldwide towards inventing new-age technologies. Our mission, as part of the research community is to bring the highest quality research to the widest possible audience. International Journal of Computer Applications is a global effort to consolidate dispersed knowledge and aggregate them in a search-able and index-able form.

The perspectives presented in the journal range from big picture analysis which address global and universal concerns, to detailed case studies which speak of localized applications of the principles and practices of computational algorithms. The journal is relevant for academics in computer science, applied sciences, the professions and education, research students, public administrators in local and state government, representatives of the private sector, trainers and industry consultants.

Indexing

International Journal of Computer Applications (IJCA) maintains high quality indexing services such as Google Scholar, CiteSeer, UlrichsWeb, DOAJ (Directory of Open Access Journals) and Scientific Commons Index, University of St. Gallen, Switzerland. The articles are also indexed with SAO/NASA ADS Physics Abstract Service supported by Harvard University and NASA, Informatics and ProQuest CSA Technology Research Database. IJCA is constantly in progress towards expanding its contents worldwide for the betterment of the scientific, research and academic communities.

Topics

International Journal of Computer Applications (IJCA) supports a wide range of topics dealing in computer science applications as: Embedded Systems, Pattern Recognition, Signal Processing, Robotics and Micro-Robotics, Theoretical Informatics, Quantum Computing, Software Testing, Computer Vision, Digital Systems, Pervasive Computing etc.

Open Review

International Journal of Computer Applications approach to peer review is open and inclusive, at the same time as it is based on the most rigorous and merit-based „blind’ peer-review processes. Our referee processes are criterion-referenced and referees selected on the basis of subject matter and disciplinary expertise. Ranking is based on clearly articulated criteria. The result is a refereeing process that is scrupulously fair in its assessments at the same time as offering a carefully structured and constructive contribution to the shape of the published paper.

Intellectual Excellence

The result is a publishing process which is without prejudice to institutional affiliation, stage in career, national origins or disciplinary perspective. If the paper is excellent, and has been systematically and independently assessed as such, it will be published. This is why International Journal of Computer Applications has so much exciting new material, much of it originating from well known research institutions but also a considerable amount of brilliantly insightful and innovative material from academics in lesser known institutions in the developing world, emerging researchers, people working in hard-to-classify interdisciplinary spaces and researchers in liberal arts colleges and teaching universities.

includes members of research center heads, faculty deans, department heads, professors, research scientists, experienced software development directors and engineers.

Dr. T. T. Al Shemmeri, Staffordshire University, UK Bhalaji N, Vels University

Dr. A.K.Banerjee, NIT, Trichy Dr. Pabitra Mohan Khilar, NIT Rourkela Amos Omondi, Teesside University Dr. Anil Upadhyay, UPTU

Dr Amr Ahmed, University of Lincoln Cheng Luo, Coppin State University Dr. Keith Leonard Mannock, University of London Harminder S. Bindra, PTU

Dr. Alexandra I. Cristea, University of Warwick Santosh K. Pandey, The Institute of CA of India Dr. V. K. Anand, Punjab University Dr. S. Abdul Khader Jilani, University of Tabuk Dr. Rakesh Mishra, University of Huddersfield Kamaljit I. Lakhtaria, Saurashtra University Dr. S.Karthik, Anna University Dr. Anirban Kundu, West Bengal University of

Technology

Amol D. Potgantwar, University of Pune Dr Pramod B Patil, RTM Nagpur University Dr. Neeraj Kumar Nehra, SMVD University Dr. Debasis Giri, WBUT

Dr. Rajesh Kumar, National University of Singapore Deo Prakash, Shri Mata Vaishno Devi University

Dr. Sabnam Sengupta, WBUT Rakesh Lingappa, VTU

D. Jude Hemanth, Karunya University P. Vasant, University Teknologi Petornas

Dr. A.Govardhan, JNTU Yuanfeng Jin, YanBian University

Dr. R. Ponnusamy, Vinayaga Missions University Rajesh K Shukla, RGPV

Dr. Yogeshwar Kosta, CHARUSAT Dr.S.Radha Rammohan, D.G. of Technological Education

T.N.Shankar, JNTU Prof. Hari Mohan Pandey, NMIMS University

Dayashankar Singh, UPTU Prof. Kanchan Sharma, GGS Indraprastha Vishwavidyalaya

Bidyadhar Subudhi, NIT, Rourkela Dr. S. Poornachandra, Anna University Dr. Nitin S. Choubey, NMIMS Dr. R. Uma Rani, University of Madras Rongrong Ji, Harbin Institute of Technology, China Dr. V.B. Singh, University of Delhi

Anand Kumar, VTU Hemant Kumar Mahala, RGPV

Prof. S K Nanda, BPUT Prof. Debnath Bhattacharyya, Hannam University Dr. A.K. Sharma, Uttar Pradesh Technical

University

Dr A.S.Prasad, Andhra University Rajeshree D. Raut, RTM, Nagpur University Deepak Joshi, Hannam University Dr. Vijay H. Mankar, Nagpur University Dr. P K Singh, U P Technical University Atul Sajjanhar, Deakin University RK Tiwari, U P Technical University

Navneet Tiwari, RGPV Dr. Himanshu Aggarwal, Punjabi University

Ashraf Bany Mohammed, Petra University Dr. K.D. Verma, S.V. College of PG Studies & Research Totok R Biyanto, Sepuluh Nopember R.Amirtharajan, SASTRA University

Sheti Mahendra A, Dr. B A Marathwada University Md. Rajibul Islam, University Technology Malaysia Koushik Majumder, WBUT S.Hariharan, B.S. Abdur Rahman University Dr.R.Geetharamani, Anna University Dr.S.Sasikumar, HCET

Rupali Bhardwaj, UPTU Dakshina Ranjan Kisku, WBUT

Gaurav Kumar, Punjab Technical University A.K.Verma, TERI Prof. B.Nagarajan, Anna University Vikas Singla, PTU

Dr H N Suma, VTU Dr. Udai Shanker, UPTU

Anu Suneja, Maharshi Markandeshwar University Prof. Rachit Garg, GNDU

Aung Kyaw Oo, DSA, Myanmar Dr Lefteris Gortzis, University of Patras, Greece. Suhas J Manangi, Microsoft Mahdi Jampour, Kerman Institute of Higher Education Prof. D S Suresh, Pune University Prof.M.V.Deshpande, University of Mumbai

Prof. Surendra Rahamatkar, VIT Prof. Shishir K. Shandilya, RGPV

M.Azath, Anna University Liladhar R Rewatkar, RTM Nagpur University R. Jagadeesh K, Anna University Amit Rathi, Jaypee University

Dr. Dilip Mali, Mekelle University, Ethiopia. Dr. Paresh Virparia, Sardar Patel University Morteza S. Kamarposhti , Islamic Azad University

of Firoozkuh, Iran

Dr. D. Gunaseelan Directorate of Technological Education, Oman

Dr. M. Azzouzi, ZA University of Djelfa, Algeria. Dr. Dhananjay Kumar, Anna University

Jayant shukla, RGPV Prof. Yuvaraju B N, VTU

Dr. Ananya Kanjilal, WBUT Daminni Grover, IILM Institute for Higher Education Vishal Gour, Govt. Engineering College Monit Kapoor, M.M University

Dr. Binod Kumar, ISTAR Amit Kumar, Nanjing Forestry University, China. Dr.Mallikarjun Hangarge, Gulbarga University Gursharanjeet Singh, LPU

Dr. R.Muthuraj, PSNACET Mohd.Muqeem, Integral University

Dr. Chitra. A. Dhawale, Symbiosis Institute of Computer Studies and Research

Dr.Abdul Jalil M. Khalaf, University of Kufa, IRAQ.

Dr. Rizwan Beg, UPTU R.Indra Gandhi, Anna University

V.B Kirubanand, Bharathiar University Mohammad Ghulam Ali, IIT, Kharagpur Dr. D.I. George A., Jamal Mohamed College Kunjal B.Mankad, ISTAR

Raman Kumar, PTU Lei Wu, University of Houston – Clear Lake, Texas. G. Appasami , Anna University S.Vijayalakshmi, VIT University

Dr. Gurpreet Singh Josan, PTU Dr. Seema Shah, IIIT, Allahabad Dr. Wichian Sittiprapaporn, Mahasarakham

University, Thailand.

Chakresh Kumar, MRI University, India

Dr. Vishal Goyal, Punjabi University, India Dr. A.V.Senthil Kumar, Bharathiar University, India R.C.Tripathi, IIIT-Allahabad, India Prof. R.K. Narayan , B.I.T. Mesra, India

Hybrid SWOT-AHP Analysis of Saudi Arabia E-Government Authors : Saleh Alshomrani, Shahzad Qamar

1-7 Sign Language Recognition in Robot Teleoperation using Centroid Distance Fourier Descriptors

Authors : Rayi Yanu Tara, Paulus Insap Santosa, Teguh Bharata Adji

8-12 Energy Saving and Survival Routing Protocol for Mobile Ad Hoc Networks

Authors : Baisakh, Nileshkumar R. Patel

13-17 Combination of Adaptive Resonance Theory 2 and RFM Model for Customer Segmentation in Retail

Company

Authors : I Ketut Gede Darma Putra, A. A. Kt. Agung Cahyawan, Dian Shavitri H.

18-23

A Novel Bi-directional Series Parallel Resonant Converter for Power Factor Correction Authors : Sivachidambaranathan. V, Subhransu Sekhar Dash

24-29 Combining Chaotic Encryption and Frequency Watermarking for Prior and Ahead Image Securing

Authors : Hassen Seddik, Ben Braiek Ezzeddine

30-36 Shuffled Frog Leaping Algorithm based Voltage Stability Limit Improvement and Loss

Minimization Incorporating FACTS Devices under Stressed Conditions uthors : L. Jebaraj, C. Christober Asir Rajan, S. Sakthivel

37-44

Diagnosing Vulnerability of Diabetic Patients to Heart Diseases using Support Vector Machines Authors : G. Parthiban, A. Rajesh, S. K. Srivatsa

Hybrid SWOT-AHP Analysis of Saudi Arabia

E-Government

Saleh Alshomrani

Faculty of Computing and IT

King Abdulaziz University

Jeddah, Saudi Arabia

Shahzad Qamar

Faculty of Computing and IT

King Abdulaziz University

Jeddah, Saudi Arabia

ABSTRACT

E-government has become an important phenomenon which attracted every country toward itself. However e-government is facing challenges which have significant effects on its performance. In order to overcome these challenges good strategies are inevitable for the e-government implementation.

In this paper we analyzed Saudi Arabia’s e-government using strengths, weaknesses, opportunities and threats (SWOT) technique and identified the intensities of SWOT factors using analytic hierarchy process (AHP) technique. The prioritized SWOT factors are then used to formulate alternative implementation strategies using TOWS matrix. The results indicated that user centric strategy and bridging digital divide strategy followed by good communication strategy and citizen awareness strategy are the best strategies that should be implemented for successful implementation of e-government in Saudi Arabia.

Keywords

SWOT, AHP, TOWS, E-government

1.

INTRODUCTION

E-government [1] is not a new concept any more. After the successful implementation of the e-commerce, governments were under pressure to provide the public services using modern telecommunication technologies. E-government has been recognized as an important tool for providing public services effectively and efficiently not only to the citizens (G2C) but also to the businesses (G2B), government employees (G2E) and other governments (G2G). E-government utilize modern ICT technologies like internet, World Wide Web (WWW) and mobile technologies for better delivery of public services, improved interactions with businesses and industries, citizen participation and more efficient government management [2] which results in less corruption, increased transparency, revenue growth and reduce cost [3]. Along with these advantages, e-government is facing several political, technical,

Economical and social challenges [4] which restrict this prominent concept to be successfully implemented and if ignored it can results in wastage of government and stakeholders resources. Therefore keeping the challenges faced by the e-government in mind there should be simple, clear and understandable strategies before implementation of the system. E-government strategies are plan for government system and their supporting infrastructure to maximize the ability of the top level management to achieve organization objectives [5].

Strengths, Weaknesses, Opportunities and Threats (SWOT) is a well known strategic planning tool to evaluate the internal

strengths and weaknesses and external opportunities and threats related to the product, service, system and organization [6]. The concept of the SWOT is shown in table 1.

Table 1: SWOT Framework Internal Factors External Factors

Strengths

Available resources which can be effectively used to achieve the objectives.

Opportunities

Favorable situation in the external environment

Weaknesses

Limitations and faults that makes achieving

objectives difficult.

Threats

Unfavorable situation in the external environment

TOWS matrix was developed by Weihrich in 1982 as the next step of SWOT for developing alternative strategies [7]. TOWS matrix provides means to develop strategies based on logical combination of factors related to internal strengths or weaknesses with factor related to external opportunities or threats. TOWS matrix identifies four conceptually distinct strategic groups shown in table 2.

Table 2: TOWS Matrix Internal Strengths

(S)

Internal Weaknesses

(W) External

Opportunities (O)

SO: Maxi-Maxi Strategy

WO: Mini Maxi Strategy

External Threats

(T)

ST: Maxi Mini Strategy

WT: Mini Mini Strategy

Analytic hierarchy Process (AHP) is multi-criteria decision making tool which uses hierarchical structure formation to show the problem and then perform pair wise comparison between the factors in order to prioritize them using Eigen-value calculation framework [8]. The information derived from pair wise comparison can be shown in a reciprocal

matrix of weights, where the assigned relative weight enters into the matrix as an element and reciprocal of the entry goes to the opposite side of the main diagonal as shown in equation 1.

�= � =

1 1 1 2 … 1 �

2 1 2 2 …

. . . . . . . . . _� 1 � 2 … � �

(1)

The rows in matrix A indicate ratios of weights of each factor with respect to all others (Eq. (1)). In the matrix when i=j, then aij = 1. Consistency should be checked for the above matrix A using the formula.

��= (� _� − �) (� −1) (2)

��= �� �� (3)

CI is the consistency index and RI is random index. The matrix is considered consistent if CR ≤ 0.1 [8].

The purpose of this study is to investigate the combined usage of SWOT and AHP as analytical process for Saudi Arabia e-government strategic planning to overcome challenges faced by the Saudi e-government. In this study, factors that affecting Saudi e-government are determined and examined with SWOT analysis method and weighting of the factors are determined by AHP method.

The rest of the paper is organized as follows. The next section presents literature review of the concept upon which this research is based. Section III presents the methodology used in this research. Section IV discusses results of SWOT-AHP application for Saudi Arabia. The last section concludes the paper.

2.

LITERATURE REVIEW

SWOT (Strengths, Weaknesses, opportunities and Threats) is commonly used situational assessment method [9] which identifies the internal strengths and weaknesses and highlights external opportunities and threats to the product, technology, planning or management. The SWOT analysis popular framework and gained acceptance because of its simplicity and power for strategy development [10]. Researcher used SWOT analysis method in different areas. In [11], the authors applied SWOT analysis technique to e-government in Uttarakhand, India and argued good governance improve the process of decision making and the process by which decisions are implemented. In [12] and [13] the authors reviewed and evaluated the vision, objectives and strategic framework of e-government in Singapore and Iran using SWOT analysis respectively. The author in [14] has applied the SWOT analysis method to evaluate the e-government implementation in Ghana. Some researchers argued that SWOT analysis technique is oversimplified [15] and has a limitation that the importance of the identified factors in decision making cannot be measured quantitatively [16] therefore SWOT alone is insufficient for decision making. In this study Analytic Hierarchy Process (AHP) is combined with SWOT analysis which provides a quantitative measure of importance of identified factors on decision making.

The Analytic Hierarchy Process (AHP) is well known multi-criteria decision making method developed by Thomas Saaty

in 1970. AHP uses hierarchal structure to show the problem and then develop priorities for alternative based on the decision of the user [8]. The application of AHP is common in economic, social, military and management science [17]. As mentioned before, SWOT analysis is a simple and effective method to identify the internal and external factors but cannot measure the intensities of the identified factors. By combining SWOT with AHP it will be easy to evaluate SWOT factors and equate the intensities [18]. Few studies previously used SWOT-AHP combined model to evaluate the e-government. In [19], the authors carried out SWOT-AHP analysis to evaluate e-government stage model. In [20], the author evaluated the e-government of Turkey using SWOT and AHP combined model.

The mentioned literature dealt with prioritization of the SWOT factors, strategies were not included based on prioritized SWOT factors. In this study SWOT-AHP combined model is used to prioritize the internal and external factors and then followed by developing alternative strategies based on those prioritized factors in the form of TOWS matrix.

3.

METHODOLOGY

The methodology adopted for this research consists of the following four steps.

3.1

SWOT Analysis

The first step of the methodology consists of SWOT analysis. SWOT analysis is employed in this study to identify strengths (S), weaknesses (W), opportunities (O) and threats (T) of Saudi Arabia e-government. The SWOT factors are identified from the existing literature available on Saudi Arabia e-government.

Table 3 illustrates the SWOT analysis which identifies the strengths, weaknesses, opportunities and threats to e-government of Saudi Arabia.

Table 3: SWOT Analysis for Saudi Arabia e-government STRENGTHS (S) OPPORTUNITIES (O) S1: Political wiliness and

Public Policy [21]

S2: Citizens focused

policy [22]

S3: Good ICT

Infrastructure in Saudi Arabia [23]

S4: E-Government portal

and sub-portals availability [24]

O1: Strong Economy of

Saudi Arabia [25]

O2: Potential in growth in

ICT Infrastructure [23]

O3: Legal Framework i.e.

e-transaction law and IT Criminal Law [26]

O4: Participation of

academics to support ICT [27]

O5: Better opportunities of

employment for IT professionals [28]

WEAKNESSES (W) THREATS (T) W1: Lack of IT skills [23]

W2: Digital divide [29]

problems

W3: Common culture on

e-transactions [30]

T1: De-centralized

Internet Governance [20]

T2: Individual attitude and

social culture [31]

T3: Privacy and Security

of personal information [32]

T4: Use of mobile technology [22]

3.2

Hierarchical Structure

The four levels hierarchical structure used in this study is shown in figure 1. The top level is the goal of analysis i.e. the evaluation of the Saudi Arabia e-government strategies. The second level is constituted by the evaluation strategies or e-government priority issues which need to be compared. In that context, three Saudi e-government priority issues are; SO1: Effective communication with all e-government stacks holders to increase citizen satisfaction. SO2: Education and Development and SO3: Value Management. Third level of the hierarchy is constituted by the four groups of factors as defined by the SWOT analysis technique: Strengths (S), Weaknesses (W), Opportunities (O) and Threats (T) and the lowest level is constituted by the factors included in each one

of the four groups of the previous level. Figure 1: Hierarchical Structure to Prioritize the SWOT factors of Saudi Arabia e-government

3.3

Pair wise Comparison

There is no standard way to evaluate pair wise comparison. In this study, the authors made a discussion group consist of the people who have the knowledge of e-government and strategic management. The group members discussed and compared two factors. There was a question of which of the two factors has a greater weight in the choice and how much greater, was the main question in the discussion. AHP then transforms each preference to a numerical value which can be compared and evaluated. The relative importance is given a value on a scale of 1–9 [8]. Pair-wise comparisons were made separately for each set of the hierarchy. For example, pair-wise comparisons of factors within each SWOT group are needed. The number of pair wise comparisons is dependent on the number of factors within the same hierarchy level. If there are n factors, the number of comparisons within the level are required based on the equation: � � −1) /2.

O 4 O 3

Goal (G)

Significant Objectives (SO)

SO1 SO2 SO3

Strengths (S)

Weaknesse s (W)

Opportunit ies (O)

Threats (T)

S 1

S 2

S 3

S 4

W

1

W 3 W 2

O 2 O 1

T 4 T 2

T 3 T 1

O 5

Table 4: Factor priority scores and global priority scores for SWOT factors SWOT GROUP SCALING

FACTOR

SWOT FEATURES LOCAL PRIORITY

GLOBAL PRIORITY

Strengths (S) 0.49

S1: Political Willingness S2: Citizen focused policy S3: Good ICT Infrastructure S4: E-Government Portals availability

0.68 [1]

0.14 [2]

0.10 [3]

0.07 [4]

0.333 0.068 0.049 0.034 λmax = 4.208 CI = 0.0693 CR = 0.0770 Weaknesses (W) 0.06 W1: Lack of IT Skills W2: Digital Divide W3: Common culture on e-transactions 0.08 [3]

0.47 [1]

0.13 [2]

0.004 0.028 0.007 λmax = 3.11 CI = 0.06 CR = 0.096 Opportunities (O) 0.36 O1: Strong Economy O2: Potential growth in ICT O3: Legal Framework for e-gov O4: Participation of academics O5: Better employment opportunities for IT professionals 0.18 [3]

0.41 [1]

0.06 [4]

0.04 [5]

0.31 [2]

0.0648 0.1476 0.0216 0.0144 0.1116 λmax = 5.36 CI = 0.09 CR = 0.08 Threats (T) 0.10 T1: De-Centralized Internet Gov T2: Individual attitude T3: Privacy and Security T4: Use of Mobile technology 0.12 [3]

0.63 [1]

0.06 [4]

0.20 [2]

0.012 0.063 0.006 0.02 λmax = 4.26 CI = 0.0889 CR = 0.098

3.4

Strategy formulation using TOWS

Matrix

A successful implementation of the e-government required a comprehensive strategy which is benched marked on global best practices and also applicable to country particular political, economic and social conditions. Several strategies can be proposed for e-government, but the selection or adoption of best strategies is important. This can easily be done by using combined SWOT-AHP method. In order to draw out best strategies, the SWOT table has to be searched for logical combinations. The formulation of those alternative strategies starts with finding those combinations. The TOWS matrix draws four logical combinations (strategies); First, SO-strategies, secondly WO-SO-strategies, thirdly, ST-strategies and fourthly WT-strategies. Table 5 shows the TOWS matrix for Saudi Arabia e-government.

Table 5: TOWS Matrix for Saudi Arabia e-government STRENGHTS WEAKNESSES O

P P O R T U N I T Y

SO Strategies: Maxi-Maxi

Strategy-1: User

Centric Strategy (S2/S4/03/04)

Strategy -2: Legal

framework for e- governance. (S2/03)

WO Strategies: Mini-Maxi

Strategy -3: Bridging

Digital Divide Strategy (W2/W4/O4/O5)

Strategy -4: Human

Capacity Building Strategy (W1/O3/O4)

T H R E A T S

ST strategies: Maxi-Mini

Strategy -5: Centralized

e- government System Strategy (S1/S3//T1/T3)

Strategy -6: Pro-active

Communication Strategy (S3/S4/T3)

WT Strategies: Mini-Mini

Strategy -7: Citizen

Awareness Strategy (W2/W3//T2/T3)

Strategy -8: Internet and

PC Penetration Strategy (W1/W2/T2/T4)

4.

DISCUSSION

This study intended to introduce a simple, acceptable and systematic methodology for e-government of Saudi Arabia. The ultimate success of e-government is dependent on the accuracy of effective SWOT analysis. But the problem in SWOT analysis is that it does not analytically determine the importance of factors. In order to eliminate this drawback, SWOT is combined with AHP which give priorities to the SWOT factors and compared them pair wise. In this study the two techniques were combined to analyze the Saudi Arabia e-government and developed strategies e-e-government in Saudi Arabia.

It can be seen that the values of Saudi Arabia e-government strengths and opportunities are higher than weaknesses and threats. Saudi Arabia e-government strengths are 7.04 times more important than weaknesses (0.480/0.068 = 7.04) and 6.07 times more important than threats (0.480/0.079 = 6.07). Similarly Saudi e-government opportunities are 8.220 times more important than weaknesses and 6.946 times than threats. If we compare the Saudi e-government strengths and opportunities, it is clear from analysis that both factors have almost same importance. Many opportunities are there for Saudi e-government to improve and provide best possible services to citizens and businesses. The data from table 4 is represented by the figure 2.

Figure 2: Pair wise comparison of SWOT features with respect to significant objectives

In Table 5, we suggested some strategies for Saudi e-government. User centric strategy and bridging digital divide strategy followed by good communication strategy and citizen awareness strategy should have adopted by Saudi Arabia for successful implementation and adoption of e-government. The ultimate aim of the e-government implementation is to provide best possible services to citizens therefore Saudi government should give prime importance to user centric strategy. The advantage of adopting this strategy would be that the citizens will get governmental services any time without any physical location restriction; it will be economical both for citizen and government; citizens will get efficient services and the whole system will be transparent.

Another important strategy that should be adopted by the Saudi Arabia is bridging digital divide. Digital divide is a social issue which is linked to the difference in the level of information between citizens [32]. Some of the factors which cause digital divide in Saudi Arabia are unequal access to information, lack of proper ICT infrastructure in rural areas, senior citizens who have no or low knowledge of modern technologies and low adoption of technology. The strategy for bridging or narrowing the digital divide in Saudi Arabia should include the characteristics like; to provide equal information to the citizen in non-discriminative fashion, develop the ICT infrastructure in the rural areas of the

country, to educate the population especially of the rural areas and to promote ICT skills especially in old age and uneducated people.

Good communication strategy is also must for the success of the e-government. Because of this strategy citizens will be aware of the new ways of getting the governmental services online. The social culture and individual attitude are the two most important threats to the Saudi e-government. This strategy will be helpful in eliminating the negative perception from the minds of people. Citizen awareness is also related to good communication strategy. Saudi government should encourage the people to use the online services and should adopt different means like print, radio and television advertisements to make the citizens aware of the new system. Security and privacy of personal information and people trust also play an important role in the success of the e-government adoption. Lack of trust in the e-government is the severe hindrance to its growth. Saudi Arabia government should develop a good legal framework for the e-government. A comprehensive legal framework should cover the overall aspects of the e-government from the delivery of service and provision of information to business process re-engineering within the different levels of government and its institutions.

5.

CONCLUSION

For successful implementation and adoption of e-government projects, good strategies are inevitable. Without comprehensive strategies it is difficult to get maximum benefits from the e-government system. In this research we thoroughly studied the e-government in Saudi Arabia. Using SWOT technique, a situational analysis was conducted in order to find out the strengths, weaknesses, opportunities and threats to the e-government of Saudi Arabia. The outcome of the SWOT technique was qualitative and subjective therefore we could not identify that which factors were more important than others. In order to solve this problem the SWOT technique was integrated with AHP technique in order to find quantitative strengths, weaknesses, opportunities and threats for the Saudi Arabia e-government. Based on the numeric results of the SWOT-AHP integration we developed strategies using TOWS method for Saudi Arabia e-government which will support better decision making by the higher authorities of Saudi Arabia e-government project.

In this research the strategies for the Saudi e-government were identified but did not evaluated to find out which strategies should be given importance and should be implemented first. In the future work we can evaluate these strategies using technique like Quantitative Strategies Planning Matrix (QSPM), in order to find the importance of individual strategy quantitatively.

6.

REFERENCES

[1] UN government Survey 2010: Leveraging e-government at the time of financial and economic crisis.

Available from:

http://unpan1.un.org/intradoc/groups/public/documents/u n-dpadm/unpan038853.pdf [Last Accessed on 15th March, 2012]

[2] Government of India. “Promoting e-governance: The

smart way forward”. Available online at:

http://arc.gov.in/11threp/ARC_11thReport_Ch2.pdf [Last Accessed 15th March, 2012]

[3] World Bank e-government.. http://go.worldbank.org/M1JHE0Z280 [Last Accessed on 15th march, 2012]

[4] M. Al-Sebie and Z. Irani: , 2005.,Technical and organisational challenges facing transactional e-government systems: an empirical study, Electronic Government, an International Journal 2005 - Vol. 2, No.3 pp. 247 - 276.

[5] A. Rabaiah and E. Vandijck. 2009. A Strategic Framework of e-Government: Generic and Best Practice, Electronic Journal of e-Government Volume 7 Issue 3, pp. 241 – 258,

[6] A. Avny. 2007. SWOT analysis of e-Government / Annals of University of Bucharest, Economic and Administrative Series, Nr. 1, pp. 43-54.

[7] H. Weihrich. 1982. The TOWS matix- A tool for situational analysis. Long Range Planning. Vol 15, 2. pp. 54-66.

[8] T.L. Saaty. 2008. Decision making with the analytic hierarchy process. International Journal of Service Sciences. Vol 1, No 1.

[9] H.H. Cheng and W.C. Huang. 2006. Application of

quantification SWOT analytical method”. Mathematical

and Computer Modelling. Vol 43, 1-2. pp. 158-169. [10] D. W. Pickton and S. Wright. 1998. What's swot in

strategic analysis?. Strategic Change, vol 7, pp. 101-109. [11] K. Vaisla and M. Bisht. 2010. SWOT Analysis of e-Initiative in Uttarakhand. International Journal of Computer Applications, Volume 12– No.5, pp. 5-14. [12] H. Ha ., K. Coghill. 2006. .E-Government in Singapore -

A SWOT and PEST Analysis, Asia-Pacific Social Science Review Vol.6 (2), pp.103-130.

[13] F. Rezazadeh. 2011. A SWOT and PEST analysis of e-government in Iran, in proc. of 5thSymposim on advances in science and technology.

[14] F. K. Andoh-Baidoo, and L. Agyepong. 2011 Examining the Preparedness of an Emerging Economy Towards E-Government Implementations: SWOT Analysis. South West Decision Sciences Institute Conference, Houston, TX.

[15] E. K. Valentin. 2005. Away With SWOT Analysis: Use Defensive/Offensive Evaluation Instead. The Journal of Applied Business Research. Vol 21, No 2.

[16] L. Zhaohui. 2011. Marketing Outsourcing of Chinese Sports Clubs, in proc. of IEEE 4th International Conference on Business Intelligence and Financial Engineering.

[17] J. Yang and P. Shi. 2002. “Applying Analytic Hierarchy

process in Firm’s overall performance evolution: A case study of China”, International Journal of Business. Vol 7, Issue 1.

[18] V. Wickramasinghe and S. Takano. 2009 Application of Combined SWOT and Analytic Hierarchy Process (AHP) for Tourism Revival Strategic Marketing

Planning: A Case of Sri Lanka Tourism. Journal of the Eastern Asia Society for Transportation Studies, Vol.8, [19] S. M. Shareef. Analysis of the e-Government stage

model evaluation using SWOT-AHP method”. Available from: roar.uel.ac.uk/jspui/handle/10552/1314 [Last Accessed on 20th February 2012]

[20] Kahraman, C., Demirel, N. C., and Demirel, T. 2007. Prioritization of e-Government strategies using a SWOT-AHP analysis: the case of Turkey, European Journal of Information Systems, Vol. 16, No.3, pp. 284-298. [21] A. Al-Solbi and S. Al-Harbi. 2008. An exploratory study

of factors determining e-government success in Saudi Arabia. Communications of the IBIMA. Volume 4. [22] A. Abanumy and P. Mayhew. 2005. M-government

Implications For E-Government In Developing Countries: The Case of Saudi Arabia, in the proc. of Mobile Government Consortium International.

[23] M. Alshehri and S. Drew. 2010. Challenges of e-Government Services Adoption in Saudi Arabia from an e-Ready Citizen Perspective, World Academy of Science, Engineering and Technology, vol 66.

[24] H. A. Al-Nuaim, 2011. An Evaluation Framework for Saudi E-Government, Journal of e-Government Studies and Best Practices, . Available from: www.ibimapublishing.com/journals/JEGSBP/2011/.../82 0912.pdf [Last Accessed: 20th February 2012]

[25] International Monitory Fund (IMF), http://www.imf.org/external/pubs/ft/survey/so/2011/int09 2111b.htm [Last Accesses 10th March, 2012]

[26] Yesser Program Saudi Arabia. www.yesser.gov.sa [Last Access: 20th February 2012]

[27] Cisco Systems Inc, “Saudi Arabia University Deploys

Advanced IT to Support New Education Model”.

Available from:

http://www.cisco.com/en/US/prod/collateral/switches/ps 5718/ps708/32023_PMU.pdf [Last Accessed: 22th February 2012]

[28] Council of Foreign Relations., 2011. Saudi Arabia in the New Middle East, [Last Accessed: 22th February 2012] [29] F. Al-Sobhi and V. Weerakkody., 2010. The role of

intermediaries in facilitating e-government diffusion in Saudi Arabia, in proc. of European and Mediterranean Conference on Information Systems.

[30] Regional Profile 2009. Information Society Portal for the

ESCWA Region.

http://isper.escwa.un.org/Default.aspx?TabId=65&item= 23 [Last Accessed 10th March, 2012]

[31] M. Alanezi., A.K.Mahmood and Shuib Basari. 2011. Review of e-Government Policy and Outcomes in the Kingdom of Saudi Arabia. International Journal of Computer Science and Information Security., vol 9, No 8.

[32] J.X, Dempsey., P, Anderson. and A. Schwartz. 2003. E-government and Privacy [online]. Available: http://www.internetpolicy.net/privacy/20030523cdt.pdf [Accessed: 25th February 2012]

[33] P.K. Sarkar and M. Singh., 2008. Narrowing the digital divide: the Australian situation. International Journal on Computer Science and Information Systems. Vol. 3, No. 2, pp. 27-35.

Saleh Alshomrani is a faculty member in the Information

Systems Department at King Abdulaziz University, Saudi Arabia. He is also serving as the Vice-Dean of Faculty of Computing and Information Technology, and the Head of Computer Science Department – North Jeddah branch at King Abdulaziz University. He earned his Bachelor degree in Computer Science (BSc) from King Abdulaziz University, Saudi Arabia in 1997. He received his Master degree in Computer Science from Ohio University, USA in 2001. He did his Ph.D. in Computer Science from Kent State University, Ohio, USA in 2008 in the field of Internet and Web-based Distributed Systems. Currently he is actively working in the area of Web and Internet Systems.

Shahzad Qamar is working as a Lecturer in Faculty of

Computing and IT at King Abdulaziz University Saudi Arabia. Shahzad Qamar has done his MS in Wireless and Mobile Networks from Bournemouth University UK.

Shahzad Qamar’s research focus is on the new emerging

technologies like e-government, M-government, Green and Cloud Computing, QoS in WiMAX and WiFi Networks.

Appendix A: Pair wise comparisons matrices for SWOT factors

Strengths S1 S2 S3 S4 Local Weights S1: Political wiliness and Public Policy 1 7 7 7 0.68 S2: Citizens focused policy 1/7 1 2 2 0.14 S3: Good ICT Infrastructure in Saudi Arabia 1/7 1/2 1 2 0.10 S4: E-Government portal and sub-portals 1/7 1/2 1/2 1 0.007

availability

Total 1.43 9 10.50 12 1 λmax = 4.208 CI = 0.0693 CR = 0.0770

Weaknesses W1 W2 W3 Local Weights W1: Lack of IT skills 1 1/7 1/3 0.08 W2: Digital divide 7 1 5 0.72 W3: Common culture on e-transactions 3 1/5 1 0.19 Total 11 1.34 6.33 1 λmax = 3.11 CI = 0.06 CR = 0.09

Opportunities O1 O2 O3 O4 O5 Local weights O1: Strong Economy of Saudi Arabia 1 1/5 3 5 1 0.18

O2: Potential in growth in ICT Infrastructure 5 1 7 7 1 0.41 O3: Legal Framework 1/3 1/7 1 2 1/7 0.06 O4: Participation of academics to support ICT 1/5 1/7 1/2 1 1/7 0.04 O5: Better opportunities of employment for 1 1 7 7 1 0.31

IT professionals

Total 7.53 2.49 18.5 22 3.29 1

λmax = 5.36 CI = 0.09 CR = 0.08

Threats T1 T2 T3 T4 Local Weights T1: De-centralized Internet Governance 1 1/5 3 1/3 0.12

T2: Individual attitude and social culture 5 1 9 5 0.63 T3: Privacy and Security of personal information 1/3 1/9 1 1/3 0.06 T4: Use of mobile technology 3 1/5 3 1 0.20 Total 9.33 1.51 16 6.67 1 λmax = 4.26 CI = 0.0889 CR = 0.09

Sign Language Recognition in Robot Teleoperation

using Centroid Distance Fourier Descriptors

Rayi Yanu Tara

1

, Paulus Insap Santosa

2

, Teguh Bharata Adji

3

EE & IT Department, Gadjah Mada University

Yogyakarta, Indonesia

ABSTRACT

Commanding in robot teleportation system can be done in several ways, including the use of sign language. In this paper, the use of centroid distance Fourier descriptors as hand shape descriptor in sign language recognition from visually captured hand gesture is considered. The sign language adopts the American Sign Language finger spelling. Only static gestures in the sign language are used. To obtain hand images, depth imager is used in this research. Hand image is extracted from depth image by applying threshold operation. Centroid distance signature is constructed from the segmented hand contours as a shape signature. Fourier transformation of the centroid distance signature results in fourier descriptors of the hand shape. The fourier descriptors of hand gesture are then compared with the gesture dictionary to perform gesture recognition. The performance of the gesture recognition using different distance metrics as classifiers is investigated. The test results show that the use of 15 Fourier descriptors and Manhattan distance-based classifier achieves the best recognition rates of 95% with small computation latency about 6.0573 ms. Recognition error is occurred due to the similarity of Fourier descriptors from some gesture.

Keywords

Hand Gesture, Sign Language, Fingerspelling, CeFD, Fourier Descriptor.

1.

INTRODUCTION

Hand gesture recognition provides a natural way to communicate with machines (e.g. robot). The use of robot, especially in teleoperation, can reduce the risk factor of task failures and human harms during several activities such as hazardous material handling [1]. Sign language is often used as input method in teleoperation system. Several works that employ visually captured hand gesture in robot teleoperation system has been researched since last decade. Color camera was used to acquire hand image, and the acquired image was processed with Artificial Neural Network (ANN) [2], Fuzzy C-means clustering (FCM) [3], or Support Vector Machine (SVM) [4] to classify the meanings of each hand gesture. In a sign language, hand shape can give information of hand gestures. Recognizing gestures through hand shape is a challenging process. Some sign language has the similar hand shape, and similar hand shape can be interpreted as different sign because of different viewpoint. Due to its complexity, the research of hand gesture recognition based on hand shape is continuously performed. Recognition of hand posture using two different shape descriptors had been conducted in [5]. Fourier descriptors and hu moments were compared in this research. This experiment used 64 fourier coefficients. Two databases of gestures were used in this experiment (i.e. Triecsch database [6] and self-made database). The result of

this experiment showed that fourier descriptors give very good recognition rate rather than hu moments. Performance comparison of fourier descriptors and geometric moment invariants was presented [7]. The comparison was used ASL database to analyze discrimination and feature invariance of hand images. The results showed that both descriptors are unable to differentiate some classes in ASL. Another shape descriptor comparison in hand posture recognition from video was also presented in [8]. The research compares: 1) Hu moments, 2) Zenike moments, 3) Fourier descriptors (common set), and 4) Fourier descriptors (complete set). The recognition also evaluated the use of several classifiers to measure similarity of hand posture with the stored hand posture in the database. Bayesian classifier, support vector machine, k-nearest neighbor, and Euclidean distance were evaluated. Overall result of the presented research showed that the common set fourier descriptors has the highest recognition rate when combined with k-nearest neighbor, reaching 100% in classifiying learning set, and 88% in test set. An application of real time hand gesture recognition system using fourier descriptors was presented in [9]. The presented system uses 56 fourier descriptors from 64 fourier coefficients, and linear combination of Hidden Markov Models (HMM) and Recurrent Neural Network (RNN). With real time processing rates of 22 frames per second and 91.9% correct classification; the presented system achieved good performance. Another approach of hand gesture recognition using fourier descriptors and hidden markov models was also introduced [10]. The system was able to recognize 20 different gestures with average recognition rate of above 90%. All of the research in [5], [7], [8], [9], and [10] were used the same shape signature (i.e. complex coordinate) in fourier descriptors calculation and used hand image from color camera.

An evaluation of the use of different shape signature in fourier descriptors calculation for shape retrieval was presented [11]. The research compared four shape signatures: 1) complex coordinate, 2) centroid distance, 3) curvature signature, and 4) cumulative angular function. Euclidean distance was used as similarity measurement. To measure the effect of each shape signature in representing a shape, precision and recall ratio were used. The result showed that the use of centroid distance signature in calculating fourier descriptors is significantly better than other shape signature. The centroid distance fourier descriptors is robust and information preserving. This is due to the centroid distance, which captures both local and global features of the shape.

This paper contributes to the use of centroid distance fourier descriptors (CeFD) in sign language recognition system, which will be employed in robot teleoperation system. The sign language is adopted from American Sign Language (ASL) fingerspelling. Only static gesture sign is researched. A

depth imager is utilized to acquire hand image. The centroid distance fourier descriptors are used as the feature vector of inputted gesture. To perform recognition, the fourier descriptors of inputted hand image are compared with the fourier descriptors of each character stored in the gesture dictionary. To obtain the best recognition rate, different classifiers will be evaluated. The following section will clearly explain our research methodology.

2.

METHODOLOGY

An overview about the methodology of this research is illustrated in Figure 1.

This research employs depth imager to acquire hand image from the human signers. The use of depth imager has benefit in segmenting hand image. Rather than color-based segmentation, segmentation in depth image is more robust since the lighting variation does not affect the image quality. Contour of the segmented hand image is then used to generate centroid distance signature. Hand contour coordinates are arranged into centroid distance signature with equal arc-length point sampling, results in N sampled signature point. Fourier transform of the centroid distance signature yields fourier descriptors, which represent feature of each hand gesture. Generally, this research is separated into two phases: dictionary build phase and classification phase. In the dictionary build phase, the fourier descriptors of each character are stored into a database to develop gesture dictionary. The gesture dictionary and comparison method are derived from [12]. The classification phase has the similar step, except that the fourier descriptors are compared with the dictionary using distance metric as classification methods. The result of classification phase is the meaning of the acquired gesture sign.

3.

FEATURES EXTRACTION

3.1

Hand Segmentation

Microsoft KinectTM is utilized as depth imager in this research [13]. Human hand is assumed as the closest object in the

imager field of view. The closer the object distance to the imager, the lower the voxel (depth pixel) value is. Threshold operation is applied to the image with a threshold value. The threshold value is acquired by summing the hand depth and the closest object distance. This method is adopted from [14]. Figure 2 shows the original image and the segmented hand image.

Figure 2. Depth image (top) and the segmented hand image (bottom)

3.2

Centroid Distance Signature

Shape signature is used to represent shape contour of an object. The shape signature itself is a one-dimensional function that is derived from shape contour coordinate. Centroid distance signature is one of several types of shape signature. In this research, the segmented hand image is processed with canny edge detection to extract the hand contour. Afterward, the centroid distance signature of hand images is generated from the hand contour. Three hand shapes with same gesture and their centroid distance signature are shown in Figure 3. The centroid distance signature r(t) is computed from the coordinates of each contour sequence by applying Equation (1) and Equation (2).



 







2

1 2 2

)

(

)

(

)

(

t

x

t

x

c

y

t

y

c

r









(1)

)

(

1

),

(

1

t

y

L

y

t

x

L

x

c



c



(2)

Depth Image Acquisition

Hand Segmentation

Hand Contour Extraction

Centroid Distance Signature

Centroid Distance Fourier Descriptor Classification

Classified Gesture

Gesture Dictionary

Figure 1. Research methodology

(a) (b) (c) (d)

Figure 3. Three hand images with same gesture and its centroid distance signature (d); original (a-blue), scaled up 50%

where xc and yc are the centroid coordinate of hand shape, x(t)

and y(t) are the coordinates of each contour, L is the contour length, and t is the contour index. As shown in Figure 3, the centroid distance signature r(t) has the translation invariant property. Rotation of the hand image causes circular shift, and scaling of hand image changes the signature value linearly.

3.3

Centroid Distance Fourier Descriptors

Centroid distance Fourier Descriptors (CeFD) was empirically proven for having higher performance rather than other fourier descriptors [11], [15]. In general, the CeFD is obtained by applying fourier transform on a centroid distance signature. The discrete fourier transform of centroid distance signature r(t) is given in Equation (3).











 





 

N

nt

j

t

r

N

a

N t n



2

exp

)

(

1

1 0

, n=0, 1, …, N-1 (3) N is the total number of sampled points from the signature, and an is the fourier descriptor. This research assigns 64 as the

N value. The sampling points of centroid distance signature are obtained by applying equal-arc length sampling, which is done by dividing the total contour length L by N. Since the centroid distance signature is real value, there are only N/2 different frequencies in the fourier transform. To make the fourier descriptors invariant to scaling, rotation, and translation, Equation (4) is employed to normalize the fourier descriptors. 0 2 / 0 3 0 2 0 1

,...,

,

,

a

a

a

a

a

a

a

a

CeFD



N

, (4)

where CeFD is the normalized fourier descriptors. The descriptors use only the magnitude values since the phase values are variant to rotation. The dc-component (i.e. a0) is

used to normalize the remaining fourier descriptors to achieve scale invariant.

Figure 4 shows the fourier descriptors of three images from Figure 3.

Figure 4. Fourier descriptors of three images in Figure 3

According to the illustration, the normalized fourier descriptors of each character have small deviation. The normalized FDs are proven invariant to translation, rotation, and scaling. When used in shape retrieval, the retrieval precision degrades when using 10 FDs and does not improve significantly when using 15 FDs [16]. Thus, only the first 15 fourier descriptors is employed in this research.

4.

GESTURES CLASSIFICATION

4.1

Gesture Dictionary

Five gestures are employed as gesture vocabulary. These gestures are adopted from ASL fingerspelling. The character similarity graph in [7] is considered when choosing each character, which will be used as the five gestures. The similarity distance of FDs between each character in sign language is essential since it can reduce the possibility of false recognition. Table 1 shows the gesture vocabulary. Fingerspelling illustration in the gesture vocabulary is obtained from [17].

Table 1. Gesture vocabulary

Fingerspelling Meaning

Turn Right

Turn Left

Forward

Stop

Backward

In the gesture dictionary, each character has 15 fourier descriptors as features. To develop reliable gesture dictionary, each character is represented by five signers. Each of signer gives two hand poses with different conditions. The total training dataset is 50 gestures, 10 gestures for each character. Having 10 variations of gesture data for each character is enough since the fourier descriptors themselves have the invariant property. The dictionary GD is represented in a matrix, as shown in Equation (5).



































ij i i j

FD

FD

C

FD

C

FD

C

FD

FD

FD

C

GD

...

...

...

...

...

...

...

...

...

...

...

...

...

...

1 31 3 21 2 1 12 11 1

, (5)

where FDij is the jth fourier descriptors of ith fingerspelling

gestures, and Ci is the character of ith fingerspelling gesture.

Having 50 gestures as dictionary is not the most efficient representation and increases the computational load. Similarity between each gesture that has the same character is considered to overcome the inefficiency of dictionary size. The similarity measurement employs Euclidean distance, which shown in Equation (6). Distance between two gestures that has value less than 0.05 can be merged. This technique reduces the dictionary size to 80% from its original size, from 50 gestures into 40 gestures. Performance comparison

[11]Musunuri, S, Dehnavi, G, “Comparison of STATCOM, SVC, TCSC, and SSSC Performance in Steady State Voltage Stability Improvement” North American Power Symposium (NAPS), 2010.

[12]C.A.Canizares, Z.Faur, “Analysis of SVC and TCSC Controllers in Voltage Collapse”, IEEE Transactions on power systems, Vol.14, No.1, Feb 1999, pp.158-165. [13]MaysamJafari,Saeed Afsharnia,”Voltage Stability

Enhancement in Contingency Conditions using Shunt FACTS Devices”,EUROCON - The international conference on computer as a tool,Warsaw,Sep 9-12,IEEE,2007.

[14]Claudia Reis, Antonio Andrade and F.P.Maciel, “Line Stability Indices for Voltage Collapse Prediction”, IEEE Power Engineering conference, Lisbon, Portugal, March. 2009.

[15]A.Mohmed, G.B.Jasmon and S.Yusoff, “A static voltage collapse indicator using line stability factors”, Journal of industrial technology, Vol.7, No.1, pp.73 – 85, 1989. [16]M. M. Eusuff, K. E. Lansey, and F. Pasha, “Shuffled

frog-leaping algorithm: A memetic meta-heuristic for discrete optimization,” Engg. Optimization. Vol. 38, no. 2, pp. 129–154, 2006.

[17]Eusuff, M. M., and Lansey, K. E., “Optimization of Water Distribution Network Design Using the Shuffled Frog Leaping Algorithm”, Journal of Water Resources Planning and Management, 2003, Vol 129, No.3, pp. 210-225

[18]N.D.Reppen, R.R.Austria, J.A.Uhrin, M.C.Patel, A.Galatic,”Performance of methods for ranking a evaluation of voltage collapse contingencies applied to a large-scale network”, Athens Power Tech, Athens, Greece, pp.337-343, Sept.1993.

[19]G.C. Ejebe, G.D. Irisarri, S. Mokhtari, O. Obadina, P. Ristanovic,J. Tong, Methods for contingency screening and ranking for voltage stability analysis of power systems, IEEE Transactions on Power Systems, Vol.11, no.1, Feb.1996, pp.350-356.

[20]E.Vaahedi, et al “Voltage Stability Contingency Screening and Ranking”, IEEE Transactions on power systems, Vol.14, No.1, February 1999.

[21]S.Sakthivel, D.Mary, “Voltage stability limit improvement incorporating SSSC and SVC under line outage contingency condition by loss minimization”, European journal of scientific research, Vol.59, No.1, 2011, pp. 44 – 54.

8.

AUTHOR’S PROFILE

L. Jebaraj received the Degree in Electrical and Electronics Engineering from The Institution of Engineers (India), Kolkata and Masters Degree (Distn.) in Power Systems Engineering from Annamalai University, Chidambaram, India in 1999 and 2007 respectively. He is doing the Ph.D., Degree in Electrical Engineering faculty from Anna University of Technology, Tiruchirappalli, India. He is working as an Assistant Professor of Electrical and Electronics Engineering at V.R.S. College of Engineering and Technology, Villupuram, Tamil Nadu, India. His research areas of interest are Power System Optimization Techniques, Power system control, FACTS and voltage stability studies.

C.Christober Asir Rajan received the B.E. Degree (Distn.) in Electrical and Electronics Engineering and Masters Degree (Distn.) in Power Systems Engineering from Kamaraj University, Madurai, India in 1991 and 1996 respectively. He received his Ph.D Degree in Electrical Engineering faculty from Anna University, Chennai, India in 2004. He is currently working as an Associate Professor in Electrical and Electronics Engineering Department at Pondicherry Engineering College, Puducherry, India. His area of interest is power system optimization, operation, planning and control. He is a member of ISTE and the institution of engineers (India). S.Sakthivel received the Degree in Electrical and Electronics Engineering and Master Degree in Power Systems Engineering in 1999 and 2002 respectively. He is doing the Ph.D., Degree in Electrical Engineering faculty from Anna University of Technology, Coimbatore, India. He is working as an Assistant Professor of Electrical and Electronics Engineering at V.R.S.College of Engineering and Technology, Villupuram, Tamil Nadu, India. His research areas of interest are Power System control, Optimization techniques, FACTS and voltage stability improvement.

Diagnosing Vulnerability of Diabetic Patients to Heart

Diseases using Support Vector Machines

G. Parthiban

Research Scholar,

Dr. MGR Educational Research

and Institute,Maduravoyal,

Chennai, India.

A. Rajesh

Professor, Dept of CSE

C.Abdul Hakkeem College

of Engineering and Technology,

Melvishram, Vellore, India.

S. K. Srivatsa, PhD.

Sr. Professor, Dept of

E & I,

St.Joseph’s College of Engineering,

Chennai, India.

ABSTRACT

Data mining is the analysis step of the Knowledge Discovery in Databases process (KDD). While data mining and knowledge discovery in databases are frequently treated as synonyms, data mining is actually part of the knowledge discovery process. Data mining techniques are used to operate on large volumes of data to discover hidden patterns and relationships helpful in decision making. Diabetes is a chronic disease that occurs when the pancreas does not produce enough insulin, or when the body cannot effectively use the insulin it produces. Most of these systems have successfully employed Support Vector Machines for the classification purpose. On the evidence of this we too have used SVM classifier using radial basis function kernel for our experimentation. The results of our proposed system were quite good. The system exhibited good accuracy in predicting the vulnerability of diabetic patients to heart diseases.

Keywords

Data Mining, Diabetes, Heart Diseases, Knowledge Discovery, Support Vector Machines.

1. INTRODUCTION

Knowledge discovery in databases (KDD) also termed as Data mining aims to find useful information from large collection of data. This process consists of iterative sequence of data cleaning, data selection, data mining pattern recognition and knowledge presentation. Data mining technology is useful for extracting non trivial information from medical databases. [1], [2] It is a interdisciplinary field closely connected to data warehousing, statistics, machine learning, and neural networks. Data mining is a powerful technology with great potential to help organizations focus on the most important information in their data warehouses [3]. Data mining tools predict future trends and behaviours, help organizations to make proactive knowledge-driven decisions [4]. There are various data mining techniques available with their suitability dependent on the domain application. Data mining application in health can have tremendous potential and usefulness. It automates the process of finding predictive information in large databases. The classification model used training data set to build classification prediction model and testing data used for testing the classification efficiency.

The term "diabetes mellitus" describes a metabolic disorder of multiple aetiology characterized by chronic hyperglycaemia with disturbances of carbohydrate, fat and protein metabolism resulting from defects in insulin secretion, insulin action, or both. Diabetes is classified into two main types. Type 1 (T1B) diabetes is due to deficient insulin production. It develops during childhood and adolescence. In this case, patients require lifelong insulin injection for survival. Type 2 (T2B) diabetes is

due to body’s ineffective use of insulin. [5] Diabetes is a chronic disease which causes serious health complications including heart disease, kidney failure and blindness. [5], [6]

Heart disease is a term for variety of disease that affecting the heart such as chest pain, shortness of breath, heart attack and other symptoms. It encompasses the diverse diseases that affect the heart. [29] Chest pains arise when the blood received by the heart muscles is inadequate.

It is the more common type – accounts for 90% of diabetic cases worldwide. It develops during adulthood. It is related to obesity, lack of physical activity and unhealthy diets. Treatment involves lifestyle changes, weight loss, or oral medications or even insulin injection in some cases.

Hyperglycemia in the long term may cause damage to eyes (leading to blindness), damage to kidneys (leading to impotence and foot disorders/ amputation), increases the risk of heart disease (stroke) and insufficiency in blood flow to legs [9]. Around 366 million people have diabetes world wide according to statistics taken in the year 2011. Also it has been projected that the people with diabetes will increase to around 552 million by the year 2030. The number of people with type 2 diabetes is increasing in every country. [7], [8]

Diabetes is a major risk factor for cardiovascular disease (disease of the heart and circulatory system). It is the main cause of death in people with diabetes (around 50%). People with type 2 diabetes are likely to die 5 to 10 years earlier than people without diabetes. Most of these deaths is due to cardiovascular disease [10]. People with type 2 diabetes are more prone to have a heart attack or stroke – twice as likely as those without diabetes [11]. It has been found that a large part of the costs attributable to type 2 diabetes is due to the treatment of cardiovascular diseases [12]. Changes in lifestyle, weight loss, dietary changes and increased physical activity can greatly reduce the risks due to cardiovascular diseases [12]. Timely detection of these people will result in reduced mortality of diabetics as well as eliminating the cost due to the treatment of cardiovascular diseases. Automatic intelligent diagnosis systems can help greatly in identifying vulnerable sections of the diabetic patients. There are several systems for diagnosis and management of diabetes [13] – [17]. However these systems are designed to predict the chances of a person getting diabetes not the vulnerability of diabetic patients to heart disease.

Likewise, there are systems to predict the chances of a person getting cardiovascular disease [20], [21]. The utility of such systems in health care has been found to be quite high [18].It has been found that Support Vector Machines (SVM’s) have been quite successfully employed in such systems [22]. Hence we have used an SVM classifier for our experimentation.

This research paper is the extension of our previous work, [30] diagnosis of heart disease for diabetic patients using Naïve bayes method. Here we are using Support vector machine and it is organized as follows in the subsequent sections – section 2 gives a brief background of support vector machines, section 3 gives our experimentation methodology, section 4 gives the results of our experiments and section 5 concludes this paper.

2.

SUPPORT

VECTOR

MACHINES

(SVM) BACKGROUND

A Support Vector Machine (SVM) is a concept in statistics and computer science for a set of related supervised learning methods that analyze data and recognize pattern introduced by Corinna Cortes and Vladimir Vapnik used for classification and regression analysis. SVM have shown good performance in a number of application areas. It constructs a hyperplane or set of hyperplanes in a high- or infinite-dimensional space, which can be used for classification, regression, or other tasks. [23] SVM’s are very much useful in data classification. SVM’s classify data by finding an optimal hyper plane separating the d – dimensional data into its two classes with a maximum interclass margin. SVM’s use so called kernel functions to cast data into a higher dimensional space where the data is separable. [24], [25] SVM is a learning machine that plots the training vectors in high dimensional space and labels each vector by its class. [28] SVM based on the principle of risk minimization which aims to, minimize the error rate. [26], [27] SVM uses a supervised learning approach for classifying data. That is, SVM produces a model based on a given training data which is then used for predicting the target values of the test data. Given a labelled training set (xi,yi), SVM require the solution of the following optimization problem to perform classification [17].

Subject to,

where,

i ≥ 0, a slack variable to allow for errors in the classification

xi – training vectors, xi Rn

- function mapping xi into a higher dimension space, C – penalty parameter of the error term (usually C>0), yi– Class label, yi {1,-1}

l

3.

EXPERIMENTATION

METHODOLOGY

The methodology described in this paper is diagnosing vulnerability of diabetic patients to heart diseases and we had collected 500 records of diabetic patients to perform the experimentation. The attributes making up each record is shown in Table 1.

Table 1. Attributes used for the diagnosis Attribut

e Role

Attribu te Name

Attribu

te Type Description

Regular Sex binomin al

Sex of the patient. Takes the following values: Male, Female

Regular Age integer Age of the patient

Regular Fam/He ri

polyno mial

Indicates whether the patient’s parents were affected by diabetes. Takes the following values: Father, Mother, Both

Regular Weight numeric Weight of the patient

Regular BP polyno

mial

Blood Pressure of the patient

Regular Fasting integer Fasting Blood Sugar

Regular PP integer Post Prondial Blood Glucose

Regular A1C numeric Glycosylated Hemoglobin Test

Regular LDL integer Low Density Lipoprotein

Regular VLDL integer Very Low Density Lipoprotein

Label Vulnera

bility nominal

Indicates the

vulnerability of the patients to heart disease. Takes the following values: High, Low

Out of the 500 records, 142 records were pertaining to patients highly vulnerable to heart diseases. The remaining 358 records were pertaining to patients less vulnerable to heart disease. Since SVM processes only numeric attributes, the nominal were converted to numeric attributes by replacing each value by a unique integer. For example, the attribute Sex values are converted as follows: Male – 1 and Female – 0. The values of the attributes were then normalized to the range 0 to 1. These records were then given as input to the SVM classifier.

SVM uses kernel functions to map the data set to a high dimensional data space for performing classification. The different types of kernel functions are as follows [17]:

Linear:

Polynomial:

d ,  > 0

1

2

3

4

Radial Basis Function

:

Sigmoid:

where , r, d are kernel parameters. The choice of the kernel depends on whether the relationship between the class labels and attributes are linear or nonlinear. For nonlinear relationships, the radial basis function (RBF) kernel has been found to be a good choice as it has lesser number of hyper parameters than other nonlinear kernels. Also RBF kernel has fewer numerical difficulties [19]. Hence we have used RBF kernel in our SVM classifier.

4. RESULT ANALYSIS

The data set used for training the classifier comprises of 500 diabetic patient records out of which 142 records are of those having heart disease (positive cases) and the remaining 358 records are of those not having heart disease (negative cases). These records after sufficient pre-processing was given as input to train the SVM classifier.

The SVM classifier was trained for different values of the RBF kernel parameters, C and . The models thus obtained for each of the values of C and  where then tested for accuracy. A good classifier should be able to exhibit high accuracy for datasets unseen rather than the training data. Hence we have used 10 fold cross validation for testing the accuracy of the classifier.

In 10-fold cross-validation, we first divide the training set into 10 subsets of equal size. Sequentially one subset is tested using the classifier trained on the remaining 9 subsets. Thus, each instance of the whole training set is predicted once so the cross-validation accuracy is the percentage of data which are correctly classified. The cross validation tests prevents overfitting problem. Based on the exhaustive trials conducted, we found that for C = 5.0 and  = 1.0 the classifier exhibited the best accuracy of 94.60%. The accuracy obtained for a few values of C and  in our trials is shown in the Table 2.

Table 2. Partial results of the trials conducted C Value  Value Accuracy of the

classifier

.

.

.

.

.

.

2 0.125 89.60%

2 0.75 92.40%

4 2.5 93.20%

4 2 93.60%

4 1.5 93.80%

4 1 94.20%

5 1 94.60%

6 1.25 94%

. .

. .

. .

The ROC curve for the classifier characteristics is shown in Fig. 1

Fig 1: ROC curve for the classifier characteristics The confusion matrix indicating the accuracy of the SVM classifier for the given data set is shown in Table 3.

Table 3. The confusion matrix of the classifier True low True high Class

precision

pred. low 355 24 93.67%

pred. high 3 118 97.52%

class recall 99.16% 83.10%

Overall accuracy: 94.60% +/- 2.01% (mikro: 94.60%)

From the results obtained, it can be seen that the classifier exhibits a very high classification accuracy i.e 94.60% overall. It also shows a very high precision for the positive class (97.52%) and also the recall of the positive class is quite good (83.10%). In the case of negative classes, the classifier exhibits high precision (93.67%) as well as high recall (99.10%).

5. CONCLUSIONS

In this paper, we have shown that it is possible to diagnose heart disease vulnerability in diabetic patients with reasonable accuracy. Classifiers of this kind can help in early detection of the vulnerability of a diabetic patient to heart disease. There by the patients can be forewarned to change their lifestyle. This will result in preventing diabetic patients from being affected by heart disease, there by resulting in low mortality rates as well as reduced cost on health for the state. SVM’s have proven to be a classification technique with excellent predictive performance and also been investigated with the help of ROC curve for both training and testing data. Hence this SVM model can be recommended for the classification of the diabetic dataset.

6. ACKNOWLEDMENTS

We are grateful to Dr.V.Shesiah, Chairman and Managing director of Dr.V.Shesiah Diabetic Research Institute, Chennai for providing an access to medical diabetic data and for his involvement in this domain.

5

7. REFERENCES

[1] J. Han Kamber, M. 2006. Data Mining: Concepts and Techniques, 2nd ed. San Francisco: Morgan Kaufman. [2] U. Fayyad, G.Piatetsky-Shapiro, and P.Smyth, “From Data

Mining to Knowledge Discovery in Databases”, AI Magazine, Vol.17, pp.37-54, 1996.

[3] Introduction to Data Mining and Knowledge Discovery, Third Edition ISBN: 1-892095-02-5, Two Crows Corporation 10500 Falls Road, Potomac, MD 20854 (U.S.A.),1999.

[4] L.A.Rose, D.T., “Discovering Knowledge in Data: An Introduction to Data Mining”, ISBN O-471-66657-2, ohn Wiley & Sons, Inc, 2005.

[5] World Health Organization. Definition and diagnosis of diabetes mellitus and intermediate hyperglycemia: Available: http://www.who. int/ diabetes/en

[6] World Health Organization. Available: http: // www.who.int/topics/ diabetes mellitus/en/

[7] International Diabetes Federation (IDF), Available: http://www.idf .org/ about-diabetes

[8] American Diabetes Association Available: http: // www.diabetes.org/

[9] NJ Morrish, Wang , Stevens , Fuller JH, Keen H. “Mortality and causes of death in the WHO Multinational Study of Vascular Disease in Diabetes”. Diabetologia 2001; 44 Suppl 2:s14 – s21 Available: http:// citeulike.org /user/mpgrayer/article/3496572 [10] R.Emslie – AM Smith Gardner ID, Morris AD.

“Vascular Complications of Diabetes”. British Medical

Journal 2000; 320:1062 – 6

Available:http://bmj.com/BMJ 2000; 320 doi: 10.1136/bmj.320.7241.1062

[11] S.M. Haffner Lehto S, Ronnemaa T, Pyoyala K, Laakso M. “Mortality from Coronary heart disease in subjects with Type 2 Diabetes and Nondiabetic Subjects with and without prior Myocardial Infarction”. The New England Journal of Medicine 1998; 339:229 – 34 Available: http:// biomedcentral.com/BMC Health Services Research 2011, 11:180 doi: 10.1186/1472-6963-11-180 [12] Diabetes with cardiovascular disease- Available: www.

idf.org/fact-sheets/diabetes-cvd

[13] Y. Huang, P. McCullagh, N. Black, and R. Harper, "Feature selection and classification model construction on type 2 diabetic patient’s data," in Lecture Notes in Artificial intelligence, vol. 3275, ICDM 2004, P. Perner, Ed. Berlin: Springer-Verlag, 2004, pp. 153-162. [14] R. Bellazzi, C. Larizza, P. Magni, S. Montani, and M.

Stefanelli, "Intelligent analysis of clinical time series: an application in the diabetes mellitus domain," Artificial Intelligence in Medicine, vol. 20 pp. 37-57, 2000.Available: http://dx.doi.org/10.1016/S0933-3657(00)00052-X

[15] R. Bellazzi, "Telemedicine and Diabetes Management: Current Challenges and Future Research Directions," Journal of Diabetes Science and Technology, vol. 2, no. 1, pp. 98-104, 2008.

[16] R. Goel, A. Misra, D. Kondal, R. M. Pandey, N. K. Vikram, J. S. Wasir, V. Dhingra, and K. Luthra,

"Identification of insulin resistance in Asian Indian adolescents:classification and regression tree (CART) and logisticregression based classification rules," Clinical Endocrinology, vol. 70 pp. 717-724, 2009. [17] K. E. Heikes, B. Arondekar, D. M. Eddy, and L.

Schlessinger, "Diabetes Risk Calculator,A simple tool for detecting undiagnosed diabetes and pre-diabetes," Diabetes Care, vol. 31, no. 5, pp. 1040-1045, 2008 [18] N. Lavrac, E. Keravnou, and B. Zupan, "Intelligent Data

Analysis in Medicine," in Encyclopedia of Computer Science and Technology, vol. 42, K. e. al., Ed. New York: Dekker, 2000, pp. 113-157.

[19] W. Kong, L. Tham, K. Y. Wong, and P.Tan, "Support vector machine approach for cancer detection using amplified fragment length polymorphism (AFLP) method," Proc. the 2nd Asia-Pacific Bioinformatics Conference (APBC2004), Dunedin, New Zealand, 2004. [20] A. Karaolis, Joseph A. Moutiris, Demetra Hadjipanayi,

and Constantinos S. Pattichis. “Assessment of the Risk Factors of Coronary Heart Events Based on Data Mining With Decision Trees”. IEEE Transaction on Information Technology in Biomedicine, Vol. 14, No. 3, May2010.Available:

http://ieeexplore.ieee.org//xpls/abs_all.jsp?arnumber=53 78501DigitalObject

Identifier: 10.1109/TITB.2009.2038906

[21] K. Srinivas, Dr. G. Raghavendra Rao, and Dr. A. Govardhan. “Analysis of Coronary Heart Disease and Prediction of Heart Attack in Coal Mining Regions Using Data Mining Techniques”. The 5th International Conference on Computer Science & Education Hefei, China. August 24–27, 2010.

[22] Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin. “A Practical Guide to Support Vector Classification”. Available: http:// www.csie .ntu.edu.tw/~cjlin.

[23] Introduction to Support Vector Machine Available: http://en. wikipedia. org/wiki/Support_vector_machine

[24] Colin Campbell and Yiming Ying, Learning with Support Vector Machines, 2011, Morgan and Claypool. Available: http://www.

morganclaypool.com/doi/abs/10.2200/S00324ED1V01Y201102 AIM010?journalCode=aim

[25] H.Barakat, Andrew P.Bradley and Mohammed Nabil H.Barakat (2009) “Intelligible Support Vector Machines for Diagnosis of Diabetes Mellitus”, IEEE Transactions on Information Technology in Bio Medicine, Volume 14, Issue 4, pp 1-7, 2009. Available: http://ieeexplore. ieee.org /xpls/ abs_ all.jsp?arnumber=5378519 Digital Object Identifier: 10.1109 /TITB. 2009.2039485 [26] N. Barakat and A.P.Bradley, “Rule Extraction from

Support Vector Machines: A Sequential Covering Approach “ IEEE Transactions on Knowledge and Data Engineering, Volume 19,no.6,pp 729-741, 2007. Available:

http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=04 161896 Digital Object Identifier no. 10.1109/TKDE.2007.1023.

[27] S.Balakrishnan, R.Narayanaswamy, N.Savarimuthu, R.Samikannu “SVM Ranking with Backward Search for Feature Selection in Type II Diabetes Databases” 2008

IEEE International Conference on Systems, Man and Cybernetics. Available: http://ieeexplore. ieee.org /xpls /abs_all.jsp?arnumber=4811692 Digital Object Identifier: 10.1109 /IC SMC.2008.4811692

[28] G.Suganya, D.Dhivya “Extracting Diagnostic rules from SVM” , Journal of Computer Applications (JCA), 2011. [29] K.Srinivas et al. / “Applications of Data Mining

Techniques in Healthcare and Prediction of Heart Attacks”, International Jounal on Computer Science and

Engineering (IJCSE) Vol..02, No.02, 2010, 250-255. Available:http://www.enggjournals.com/ijcse/doc /IJCSE10-02-02-25.pdf

[30] G.Parthiban, A.Rajesh, S.K.Srivatsa, “Diagnosis of Heart Disease for Diabetic Patients using Naïve Bayes Method”, International Journal of Computer Applications (IJCA) Volume 24-No.3, June 2011, 0975-8887.Available:

http://www.ijcaonline.org/archives/volume24/number3/2 933-3887 doi 10.5120/2933-3887