Docstoc

Development of an Expert System for Diagnosis and appropriate Medical Prescription of Heart Disease Using SVM and RBF

Document Sample
Development of an Expert System for Diagnosis and appropriate Medical Prescription of Heart Disease Using SVM and RBF Powered By Docstoc
					                                                          IJCSIS) International Journal of Computer Science and Information Security,
                                                          Vol. 8, No. 5, August 2010




      .Development of an Expert System for Diagnosis
       and appropriate Medical Prescription of Heart
               Disease Using SVM and RBF
          Shaikh Abdul                    V. D. Bhagile2,                            R. R. Manza3,                        R. J. Ramteke4
                                      2                                          3                                        4
            Hannan1,                  Lecturer, Department                       Lecturer, Dept. of                       Reader, Dept. of
  1
   Lecturer, Department               of Computer Science,                    Comp. Science and I.T.                    Computer Science,
   of Computer Science,                  Sayali College,                       Dr. B.A.M.University,                     North Maharashtra
   Vivekanand College,                    Aurangabad,                               Aurangabad,                             University,
       Aurangabad,                     Maharashtra, India.                          Maharashtra                        Jalgaon, Maharashtra,
    Maharashtra, India,               v_shinde68@yahoo.co.in                  manzaramesh@gmail.com                   rakeshramteke@yahoo.co.in
      hannan_7us@yahoo.com


                                                                           The Need for Expert Systems
Abstract---                                                                Expert systems are necessitated by the limitations associated
                                                                           with conventional human decision-making processes,
Human disease diagnosis is a complicated process and requires              including:
high level of expertise. Any attempt of developing a expert system         1. Human expertise is very scarce.
dealing with human disease diagnosis has to overcome various               2. Humans get tired from physical or mental workload.
difficulties. This paper describes a project work aiming to                3. Humans forget crucial details of a problem.
develop a expert system for diagnosing of heart disease using
                                                                           4. Humans are inconsistent in their day-to-day decisions.
neural network technique. Now a days neural network are being
used successfully in an increasing number of application areas.            5. Humans have limited working memory.
This project includes the detailed information about patient and           6. Humans are unable to comprehend large amounts of data
preprocessing was done. The Support Vector Machine (SVM)                   quickly.
and Radial Basis Function (RBF) have been applied over the data            7. Humans are unable to retain large amounts of data in
for the experiment. This research project focuses on the research          memory.
and development of a web-based clinical tool designed to improve           8. Humans are slow in recalling information stored in
the quality of the exchange of health information between health           memory.
care professionals and patients. Practitioners can also use this           9. Humans are subject to deliberate or inadvertent bias in their
web-based tool to corroborate diagnosis. The proposed system is
                                                                           actions.
experimented on various scenarios in order to evaluate it’s
performance. In all the cases, proposed system exhibits                    10. Humans can deliberately avoid decision responsibilities.
satisfactory results.                                                      11. Humans lie, hide, and die.
                                                                                     The proposed methodology uses neural network for
Keywords – Neural Network, Support Vector Machine (SVM),                   classifier. The performance of proposed methodology was
RBF (Radial Basis Function), Symptoms, Medicine.                           evaluated with two different neural network techniques.
                                                                           Moreover, we compared our result with Support Vector
                    I.       INTRODUCTION                                  Machine and Radial Basis Function with original medicines
Expert systems(ES) are artificial intelligence-base computer               provided by the doctor. We obtain 97% accuracy from the
programs that have received a great deal of attention during               experiments made on the data set containing 300 samples.
years. These programs have been used to solve an impressive                This classification is the highest so for with our data.The
array of problems in a variety of fields. The part of the expert           paper is organized as following, in Section 2, a brief overview
system that stores the knowledge is called the knowledge base.             on previous related works and in section 3, introduction of
The part that holds the specifics of the to-be-solved problem is           Support Vector Machine and Radial Basis Function is
call the global database. The part that applies the knowledge              described. Section 4, the proposed methodology and preparing
to the problem is called the inference engine. Expert systems              Data for underlying neural network. Section 5, Experimental
typically have friendly user interface to enable inexperienced
users to specify problems for the system to solve and to
                                                                           1. Shaikh Abdul Hannan-Working as lecturer in Vivekanand College,
understand the system’s conclusion.                                        Aurangabad and persuing Ph.D. from North Maharashtra University, Jalgaon
                                                                           (hannan_7us@yahoo.com)
           Although, there are many computer based diagnosis               2. V. D. Bhagile-Working as Principal in Sayali College, Aurangabad.
                                                                           (v_shinde68@yahoo.com)
systems are developed for medicine. [1-3], however the
                                                                           3. R. R. Manza - Working as Lecturer in Department of Computer Science
number of expert system for human heart diagnosis domains                  and Information Technology, Dr. B. A. M. University, Aurangabad.
are still very few [4].                                                    (manzaramesh@gmail.com)
                                                                           4. R. J. Ramteke-Working as a reader in Department of Computer Science,
                                                                           North Maharashtra University,Jalgaon. (rakeshramteke@yahoo.co.in)

                                                                     245                                  http://sites.google.com/site/ijcsis/
                                                                                                          ISSN 1947-5500
                                                          IJCSIS) International Journal of Computer Science and Information Security,
                                                          Vol. 8, No. 5, August 2010



.analysis and how the coding is done with patients as well as FCM–K-means/CHMM and ANN. These experimental results
medicine data is described. Section 6, Discussion and result               were obtained for specificity and sensitivity rates 92% and
of first five patients medicine given by the expert system and             94% for CHMM, 92% and 97.26% for FCM–K-
is compared with the original medicine. Finally, we concluded              means/CHMM), respectively. Finally, Sengur et al.
this paper in Section 7.                                                   investigated the use of principal component analysis (PCA),
                                                                           artificial immune system (AIS) and fuzzy k-NN to determine
                     II. BACKGROUND                                        the normal and abnormal heart valves from the Doppler heart
                                                                           sounds [9]. For reducing the complexity, PCA was used. In the
A.        Related Works                                                    classification stage, AIS and fuzzy k-NN were used. To
   Up to now, various classification algorithms have been                  evaluate the performance of the methodology, a comparative
employed on Turkoglu’s valvular heart disease data set and                 study was realized by using a data set containing 215 samples.
high classification accuracies have been reported in the last              The validation of the method was measured by using the
decade [5-10]. Turkoglu’s valvular heart disease data set was              sensitivity and specificity parameters; 95.9% sensitivity and
obtained from Firat Medical Center. A detailed description for             96% specificity rate was obtained. Sengur et al. also
the data set will be given in the next section.                            investigated the use of Linear Discriminant Analysis (LDA)
   The valvular heart disease data set was firstly utilized in [5]         and Adaptive neuro-fuzzy inference system (ANFIS) for
where Turkoglu et al. fulfilled an expert diagnosis system                 clinical diagnosis and recognition of heart valve disorders
which uses backpropagation artificial neural networks (BP-                 [10]. The validation of the method is measured by using the
ANN) classifier. The performance evaluation of the realized                sensitivity and specificity parameters. 95.9% sensitivity and
system was evaluated by classification accuracy and the                    94% specificity rate was obtained.
correct classification rate was about 94% for normal subjects
and 95.9% for abnormal subjects. Later, Turkoglu et al.                          III. ORGANIZATION TO EXPERT SYSTEM
suggested an intelligent system for detection of heart valve
disease based on wavelet packet neural networks (WPNN) [6].                     To diagnose the heart failure cause different popular
The reported correct classification rate was about 94% for                 methods used are MRI, Doppler and Expert System. MRI can
abnormal and normal subjects. Recently, Comak et al.                       provide clear three dimensional images of the heart. Doppler
investigated the use of least-square support vector machines               technique has gained much more interest since Satomura first
(LS-SVM) classifier for improving the performance of the                   demonstrated the application of the Doppler effect to the
Turkoglu’s proposal [7]. They intended to realize a                        measurement of blood velocity in 1959[11]. However the
comparative study. Classification rates of the examined                    factor such as calcified disease or obesity often result in a
classifiers were evaluated by ROC curves based on the terms                diagnostically unsatisfactory. Doppler techniques assessment
of sensitivity and specificity. The application results showed             and therefore, it is sometimes necessary to assess the
that according to the ROC curves, the LS-SVM classifier                    spectrogram of the Doppler Shift signals to elucidate m the
performance was comparable with ANN, but the training time                 degree of the disease [12].        Many studies have been
of LS-SVM is shorter than that of the ANN and it can always                implemented the classify Doppler signals in the pattern
converge the same solution while ANN cannot. According to                  recognition field [13-14].
these results, LS-SVM’s training time is about 13 times                         Expert system is a intelligent program which olds the
shorter than ANN’s training time. This is an important                     accumulated knowledge of one or more domain experts.
difference. Because, LS-SVMs are trained only depending on                 There are many types of expert system currently exist.
support vectors, not by whole training data set. In addition,              MYCIN expert system is used in medical field for diagnosis of
LS-SVM can overcome the overfitting much successfully than                 blood disorders. DESIGN ADVISOR is another expert
ANN.                                                                       system used in processor chip design to give advice to
   More recently, Uguz et al. performed a biomedical system                designer about component placement, minimizing chip size
based on Hidden Markov Model for clinical diagnosis and                    etc. PUFF expert system is also used in medical system for
recognition of heart valve disorders [8]. The fulfilled                    diagnosis of respiratory condition of patient. PROSPECTOR
methodology was also used the database of Turkoglu et al. In               expert system is used by geologists to identify sites for drilling
the presented study, continuous HMM (CHMM) classifier                      or mining. DENDRAL expert system is used to identify the
system was used. Single Gaussian model was preferred to                    structure of chemical compounds. LITHIAN expert system
determine emission probability. The methodology was                        gives advice to archaeologists to examiner stone tools. Expert
composed of two stages. At the first stage, the initial values of          system having three main components knowledge base,
average and standard deviation were calculated by separating               inference engine and user interact. Knowledge base is the
observation symbols into equal segments according to the                   collection of facts and rules which describe all the knowledge
state number and using observation symbols appropriate to                  about problem domain. The inference engine is used to
each segment. At the second stage, the initial values of                   choose the appropriate facts and rules to apply during user
average and standard deviation were calculated by separating               query. Where as user interface takes the user query in a
observation symbols into the clusters (FCM or K-means                      readable form and passes it to the inference engine. It then
algorithms) that have equal number of states and using                     displays the result to the user. Based on much more useful
observation symbols appropriate to the separated clusters. The             but it has some limitation like limited domain, no current
implementations of the experimental studies were carried out               updation, no system self learning, no common sense, expert
on three different classification systems such as CHMM,                    needed to setup and maintain. But even though it is used

                                                                     246                               http://sites.google.com/site/ijcsis/
                                                                                                       ISSN 1947-5500
                                                         IJCSIS) International Journal of Computer Science and Information Security,
                                                         Vol. 8, No. 5, August 2010



.world wide because they are not always available, can be used related to our work investigated the use of the SVM approach
anytime anywhere, human experts are not 100% reliable or                  to select bankruptcy predictors. They reported that SVM was
consistent. Expert may not good for explanation of decisions              competitive and outperformed other classifiers (including
and cost effective. While using the expert system some legal              neural networks and linear discriminant classifier) in terms of
and ethical issues we need to follow to set the                           generalization performance [28]. In this study, we are
responsibility.[15]                                                       interested in evaluating the performance of the SVM approach
                                                                          in the domain of heart disease in comparison with that of
    A.   Support Vector Machine :                                         Radial Basis Function in neural networks.

Support Vector Machines(SVMs) are a state of the art pattern              Let us define labeled training examples [xi, yi], an input
recognition techniques whose foundation stem from statistical             vector xi  R n a class value yi   1,1, i  1,....., l For the
learning theory. However, the scope of SVMs goes beyond                   linearly separable case, the decision rules defined by an
pattern recognition because they can also handle two more                 optimal hyperplane separating the binary decision classes is
learning problems i.e. regression estimation and density                  given as the following equation in terms of the support vectors
estimation. An SVM is a general algorithm based on
guaranteed risk bounds of statistical learning theory i.e. the so
called structural risk minimization principle. It is a learning
machine capable of implementing s set of functions that
approximate best the supervisor’s response with an expected               where Y is the outcome, yi is the class value of the training
risk bounded by the sum of the empirical risk and Vapnik –                example xi, and . represents the inner product. The vector x
Chevonenkis (VC) confidence. Recent advances in statistics,               =(x1,x2,.,xn) corresponds to an input and the vectors xi,
generalization theory, computational learning theory, machine             i=1,.,N, are the support vectors. In Eq. (1), b and  i are
learning and complexity have provided new guidelines and                  parameters that determine the hyperplane.
deep insights into the general characteristics and nature of the
model building/learning/fitting process [16]. Some researchers            For the non-linearly separable case, a high-dimensional
have pointed out that statistical and machine learning models             version of Eq. (1) is given as follows:
are not all that different conceptually [17,18]. Many of the
new computational and machine learning methods generalize
the idea of parameter estimation in statistics. Among these
new methods, Support Vector Machines have attracted most                  The function K(x,xi) is defined as the kernel function for
interest in the last few years.                                           generating the inner products to construct machines with
          Support vector machine (SVM) is a novel learning                different types of non-linear decision surfaces in the input
machine introduced first by Vapnik [19]. It is based on the               space.
Structural Risk Minimization principle from computational
learning theory. Hearst et al. [20] positioned the SVM                         B. Radial Basis Function (RBF)
algorithm at the intersection of learning theory and practice:
‘‘it contains a large class of neural nets, radial basis function                   Radial basis function (RBF) networks were
(RBF) nets, and polynomial classifiers as special cases. Yet it           introduced into the neural network literature by Broomhead
is simple enough to be analyzed mathematically, because it                and Lowe [29]. The RBF network is similar to a general feed-
can be shown to correspond to a linear method in a high                   forward neural network trained using the back-propagation
dimensional feature space nonlinearly related to input space.’’           scheme. It has three layers of neurons, namely input, hidden
In this sense, support vector machines can be a good candidate            and output. However it uses only one hidden layer, each
for combining the strengths of more theory-driven and easy to             neuron in which operates as the Gaussian transfer function, as
be analyzed conventional statistical methods and more data-               against the sigmoid function [30]
driven, distribution free and robust machine learning methods.                      Mathematically, the output y of an RBF network
          In the last few years, there have been substantial              corresponding to input x is computed by the equation;
developments in different aspects of support vector machine.                              n
These aspects include theoretical understanding, algorithmic
strategies for implementation and reallife applications. S VM
                                                                          y  f ( x)     w  || x c || w
                                                                                         i 1
                                                                                                i          i       0                (Eq.A.1)

has yielded excellent generalization performance on a wide                Where wi is the connection weight between the ith hidden
range of problems including bioinformatics [21,22,23], text               neuron and output neuron; w0 the bias. φ||x-ci|| indicates a
categorization [24], image detection [25], etc. These
                                                                          RBF which is normally Gaussian having following expression;
application domains typically have involved high-dimensional                                         n
                                                                                                         || xi  ci ||2
input space, and the good performance is also related to the
fact that SVM’s learning ability can be independent of the
                                                                           || x c i ||  exp(   
                                                                                                    i 1      2 i2
                                                                                                                        )           (Eq. A.2)

dimensionality of the feature space.
                                                                          Where ci are centers of the receptive field; and σi the widths of
The SVM approach has been applied in several financial
                                                                          the Gaussian function which indicates the selectivity of a
applications recently, mainly in the area of time series
                                                                          neuron.
prediction and classification [26,27]. A recent study closely


                                                                    247                                        http://sites.google.com/site/ijcsis/
                                                                                                               ISSN 1947-5500
                                                       IJCSIS) International Journal of Computer Science and Information Security,
                                                       Vol. 8, No. 5, August 2010



                                                                         B. Preparing Data for underlying neural network:
                                                                             The data is collected from daily OPD session while
                                                                    doctor examining the patients. The symptoms and information
                                                                    about patients details like Previous History(p1), Present
                                                                    History(p2), Personnel History(p3), Physical Examination(p4),
                                                                    Cardio Vascular System(CVS), Respiratory Rate(RS), Per
                                                                    Abdomen(PA), Central Nervous system(CVS), ECG and
                                                                    Blood Investigation(BI). The main point is ECG from which
                                                                    the patient can easily diagnose whether the patient is having
                                                                    heart problem or not.
                                                                             All 300 patients data collected regarding heart
                                                                    disease and the data are prepared in different Excel Sheets
                                                                    which contains codes of each individual disease, history and
                                                                    symptoms. In one excel file 13 sub-sheets are taken for each
                                                                    field of information such as for Previous History (p1), for
                                                                    Present History the second sub-sheet and the name is given
                                                                    (P2), for Personnel History (P3) the third sub-sheet is taken,
                                                                    like this the data collection has 13 different sub-sheets for
            Figure 2. Schematic diagram of RBF                      different fields. All the fields are taken under the supervision
     The major task of RBF network design is to determine           of the Cardiologist
centers. The easiest way to do so is to choose the centers                   The code is given to each symptoms, physical
randomly from the training set. Another approach is to use the      examination parameter or diseases in each sub-sheet for
k-means technique consisting of clustering the input training       experimental work. On this data some pre-processing i.e.
set into groups and choose the mean of each group as the            normalization, coding and decoding methods are applied for
center. Also, the centers can be treated as a network parameter     the expected output.
along with wi and adjusted through error-correction training.                In table 1, the Previous History (P1) has represented
After the center is determined, the connection weights wi           with 1 to 18 different diseases of total 300 heart patients and
between the hidden layer and output layer can be obtained           represents the codes respectively from 1 to 18. The code 1
through ordinary back-propagation-based training [30].              which represents Hypertension, Code 2 represents Diabetes
                                                                    Mallitus like this it contains 18 different diseases. Some of
                IV. METHODOLOGY                                     them are as shown in Table 1.
    A. Proposed methodology and implementation of
       with FFBP and RBF                                                    Code                Name of Disease
                                                                              1     Hypertension
                                                                              2     Diabetes Mallitus
                                                                              3     TB
                                                                              4     Bronchial Asthama
                                                                              5     Hyperthyroidism
                                                                                Table 1 : Previous History of Patients
                                                                    In table 2, Present History (P2) and the symptoms present in
                                                                    P2 are represented by Codes. The Code 1 which represents
                                                                    Chest Pain/Discomfort, Code 2 represents Retrosternal Pain
                                                                    like this it contains 29 different symptoms. Some of the
                                                                    symptoms are shown in table 2.
                                                                             Code                  Symptoms
                                                                                1     Chest Pain/Discomfort
                                                                                2     Retrosternal Pain
                                                                                3     Palpitation
                                                                                4     Breathlessness
                                                                                5     sweating
                                                                                 Table 2 : Present History of patients

                                                                    In table 3, Personnel History (P3) and the information present
                                                                    in P3 are represented by codes for different bad habits. The
                                                                    Code 1 which represents Smoking, Code 2 represents Tobacco
                                                                    like this 4 different bad habits are taken and specified by 1 to 4
                                                                    codes. Some of the personnel history parameters are given
                                                                    below.



                                                                  248                            http://sites.google.com/site/ijcsis/
                                                                                                 ISSN 1947-5500
                                                      IJCSIS) International Journal of Computer Science and Information Security,
                                                      Vol. 8, No. 5, August 2010



       Code                 Personnel History
        1        Smoking                                               In table 7, Central Nervous System (CNS) and the information
        2        Tobacco                                               present in CNS are represented by codes for different
        3        Alcohol                                               symptoms. The Code 1 which represents Consciousness, Code
        4        Nil                                                   2 represents Orientation like this 5 different symptoms are
                  Table 3 : Personnel History                          found and specified by 1 to 5 codes for each symptom. Some
In table 4, Physical Examination (P4) and the information              are as shown below in table 8.
present in P4 are represented by codes for different physical
parameters. The Code 1 which represents Consciousness,                          Code               Symptoms
Code 2 represents Orientation like these 25 different physical                   1     Consciousness
parameters and specified by 1 to 25 codes for each parameter.                    2     Orientation
Some are as shown below in table 4.                                              3     Focal Deficit
        Code              Physical Examination                                   4     No Abnormality Detected (NAD)
          1      Altered Consciousness                                           5     Restlessness
          2      Orientation                                                       Table 8 : Central Nervous System
          3      Dyspnoea
          4      Fever                                                 In table 8, Electro Cardio Gram (ECG) and the information
          5      Low Pulse Rate                                        present in ECG are represented through codes for different
                Table 4 : Physical Examination                         finding which points to different problems of heart. The Code
In table 5, Cardio Vascular System (CVS) and the information           1 which represents ST Elevation, Code 2 represents Anterior
present in CVS are represented by codes for different                  Wall like this 21 different heart findings are found and
symptoms. The Code 1 which represents Heart Sound, Code 2              specified by 1 to 21 codes for each finding. Some are as
represents Normal Heart Rate like this 8 different symptoms            shown below in table 9.
and specified by 1 to 8 codes for each symptom. Some are as
shown below in table 5.                                                         Code                   ECG Point
        Code                   Symptoms                                            1      ST Elevation
          1      Heart Sounds                                                      2      Anterior Wall
          2      Normal Heart Rate                                                 3      Antero Septal
          3      Tachycardia                                                       4      Inferior
          4      Bradycardia                                                       5      Infero Posterior
          5      Regular Heart Rhythm                                               Table 9 : Electro Cardio Gram (ECG)
               Table 5 : Cardio Vascular System                        In table 10, Blood Investigation (BI) and the information
In table 6, Respiratory System (RS) and the information                present in BI are represented through codes for blood
present in RS are represented by codes for different                   investigation. The Code 1 which represents Cardiac Enzymes
symptoms. The Code 1 which represents Breath Sound                     (High), Code 2 represents Blood Sugar Test like this 24
preserved, Code 2 represents Breath Sound Reduced like this            different investigations has found and specified by 1 to 24
5 different symptoms are found and specified as shown in               codes for each investigation in all patient. Some are as shown
table 6.                                                               below in table 10.
        Code                   Symptoms                                        Code                     Symptoms
          1      Breath Sounds Preserved                                          6      Lipid Profile normal
          2      Breath Sound Reduced                                             7      Lipid Profile Abnormal
          3      Basal Crepts                                                     8      Complete Blood Count Normal
          4      No Abnormality Detected (NAD)                                    9      Leucocytosis
          5      Ranchi                                                          10      Anaemia
                 Table 6 : Respiratory System                                           Table 10 : Blood Investigation
In table 7, Per-Abdomen (PA) and the information present in            In table 11, all the medicines names along with their codes i.e.
PA are represented by codes for different symptoms. The                MID which are prescribed by the doctor to the patients. The
Code 1 which represents Liver (Hepatomegaly), Code 2                   medicine sheet contains 52 different medicines which are
represents Spleen (Splenomegaly) like these 6 different                prescribed by the doctor in different 300 stages. Some are as
symptoms have found and specified by 1 to 6 codes for each             shown below in table 11.
symptom. Some are as shown below in table 7.
          Code                 Symptoms                                       Code                Medicine Name
             1     Liver(Hepatomegaly)                                         1       Alprazolam
             2     Spleen (Splenomegaly)                                       2       Amlodepine
             3     Free Fluid Present                                          3       Aspirin
             4     Abdominal Distension                                        4       Atenolol
             5     No Abnormality Detected (NAD)                               5       Atorvastatin
                    Table 7 : Per Abdomen                                              Table 11 : Medicine Names


                                                                 249                               http://sites.google.com/site/ijcsis/
                                                                                                   ISSN 1947-5500
                                                         IJCSIS) International Journal of Computer Science and Information Security,
                                                         Vol. 8, No. 5, August 2010




In table 12, all Patients information such as previous history(P1), P2(Present History), P3(personnel History), P4(Physical
Examination), CVS(Cardio Vascular System), RS(Respiratory System), PA(Per Abdomen), CNS(Central Nervous System),
ECG(Electrocardiography) and BI(Blood Investigation) which contains all the represented codes that are present in the
individual patients.
      Sr.      Patient                                      Symptoms and Findings
      No.       Name
                           Age          P1       P2     P3         P4      CVS RS PA CNS ECG                     BT
        1         A        55M           2    1,2,5,13 4          7,10      8      4      5      4      1,3      14
        2         B        58 M          2      1,2,8    2     7,8,13,14    8      4      5      4       2        7
       3          C        60M           8     5,7,13    4       1,6,12     8      4      5      4       9       14
        4         D        60M         1,2       4,5     4 1,2,7,13,14 3,5         3      5      4      12        4
       5          E          56F           1       15,16     4       6,9,12        8     4     5      4        10       2
                            Table 12 : collection of different details of the individual Heart Patients
In table 13, different 52 medicines were used by the doctor on total 300 patients. All the medicines are prescribed by the doctor.
In this table the medicines codes are used as the description given in the table 11.


            Patient                                                                                                                     MI
    Sr.     Name          MID     MID     MID     MID      MID       MID       MID     MID      MID      MID       MID       MID        D
    No.                    1       2       3       4        5         6         7       8        9        10        11        12        13
     1     A                2       3      5       6      14     17     19      21      23     25      26                    27,29      36
     2     B                2       3      5       6      14     16     17      21      23     25      26                     27        28
     3     C                1       5      6      14      25
     4     D                3       5      7      10      11     13     14      17      19     30
     5     E                5      14     15      19
                      Table 13 : All the Medicine codes provided by the doctor to the individual patients.

             V. EXPERIMENTAL ANALYSIS                               1,2,5 and 13 so these locations are defined by 1 (one) and all
For further training of neural network process the proposed other symptoms are 0 (zero). In such a way all the fields are
information is coded in binary form (0 or 1). If the symptom is defined. All the parameter that we consider in medical
present in the patients at particular position at that point it is prescription like Sr. No., ,age , P1, P2,P3,P4,CVS, RS, PA,
defined by one (1) and if the symptoms or disease is not CNS, ECG and BT are converted in binary number where this
present at that position it is placed by Zero (0). Suppose for is used in neural network for train the neurons for achieving
example in the field P2 (present history) there are total 29 better result.
symptoms present and the patient no 1 is having the symptom
                              The individual data of the patient no 1 is defined in binary form as :
Sr No        Age                    P1                                 P2                      P3
00000001 0110111           010000000000000000           11001000000010000000000000000          0001
       P4                            CVS         RS       PA       CNS                ECG
0000001001000000000000000 00000001 00010 000010 00010                        101000000000000000000
                BT
000000000000010000000000;
                                       Symptoms and Information Coding of the patient 1.

     Using this sequence of binary format we were not getting            placed. Due of reshuffling of the fields we got satisfactory
appropriate result. Therefore we have change the order of                result upto 97%.
fields as per suggestion of the doctor because the doctors are                For this expert system total 52 different medicine are
prescribing the medicines on the basis of the ECG and blood              prescribed by the Doctor and if the medicine is present at that
investigation. So the order of ECG is changed from field no. 9           position it is defined by one (1) and if it is absent at that
to field no. 1 and after ECG we have taken Blood                         position it is defined by Zero(0). Similarly for patient one the
Investigation and rest of the fields are same and at last age is         prescribed medicine are defined as :

               0 1 1 0 1 1 0 0 0 0 0 0 0 1 0 0 1 0 1 0 1 0 1 0 1 1 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;

                                                Medicine Coding of the patient 1.

                                                                   250                                http://sites.google.com/site/ijcsis/
                                                                                                      ISSN 1947-5500
                                                        IJCSIS) International Journal of Computer Science and Information Security,
                                                        Vol. 8, No. 5, August 2010



        MID               Patient from A to I                                   MID                  Patient from A to I
                 A   B   C D E F G              H   I                                     A    B    C D E F G                H    I
          1      1   1   1    1    1   1    1   1   0                             1       0    1    1    1 0 1 1             1    1
          2      0   0   0    0    1   0    1   1   1                             2       0    0    0    0 0 0 0             0    0
          3      1   1   1    1    1   0    1   1   1                             3       1    1    1    1 1 0 1             1    1
          4      0   0   0    0    0   0    1   0   0                             4       0    0    0    0 0 0 0             0    0
          5      1   1   1    1    1   1    1   1   1                             5       1    1    1    1 1 1 1             1    1
          6      1   1   1    1    0   1    1   0   1                             6       1    1    1    1 1 1 1             1    1
          7      0   0   0    0    0   0    0   0   0                             7       0    0    0    0 0 0 0             0    0
          8      0   0   0    0    0   0    0   0   0                             8       0    0    0    0 0 0 0             0    0
          9      0   0   0    0    0   0    0   0   0                             9       0    0    0    0 0 0 0             0    0
         10      0   0   0    0    0   0    0   0   0                            10       0    0    0    0 0 0 0             0    0
         11      0   0   1    0    0   0    0   0   0                            11       0    1    0    1 0 0 0             0    1
         12      0   0   0    0    0   0    0   0   0                            12       0    0    0    0 0 0 0             0    0
         13      0   0   0    1    0   0    0   0   0                            13       0    0    0    1 0 0 0             0    0
         14      0   0   1    1    1   0    1   1   0                            14       1    1    1    1 1 0 1             1    1
         15      0   0   0    0    0   0    0   0   0                            15       0    0    0    0 1 0 0             0    0
         16      1   1   0    0    0   1    0   0   1                            16       1    1    0    0 0 1 0             0    0
         17      1   1   0    1    0   1    1   0   1                            17       1    1    0    1 0 1 1             1    1
         18      1   1   0    0    0   0    0   0   0                            18       0    0    0    0 0 0 0             0    0
         19      1   0   0    0    0   0    0   0   0                            19       0    0    0    0 0 0 0             0    0
         20      0   0   0    0    0   0    0   0   0                            20       0    0    0    0 0 0 0             0    0
         21      1   1   1    1    0   1    0   0   0                            21       1    1    1    1 0 1 1             1    1
         22      0   0   1    1    0   0    0   0   0                            22       0    1    0    0 0 0 0             0    0
         23      1   1   1    1    0   1    0   0   0                            23       1    1    0    0 0 1 0             0    0
         24      0   0   1    0    0   0    0   0   0                            24       0    1    0    0 0 0 0             0    0
         25      1   1   1    1    0   1    0   0   1                            25       1    1    1    1 0 1 1             1    1
         26      1   1   1    1    0   0    0   0   0                            26       1    1    0    0 0 0 0             0    0
         27      1   1   1    1    0   0    0   0   0                            27       1    1    0    0 0 0 0             0    0
         28      1   1   0    1    0   0    0   0   0                            28       1    0    0    0 0 0 0             0    0
         29      1   1   0    0    0   1    0   0   1                            29       1    0    0    0 0 1 0             0    0
         30      0   0   0    0    0   0    0   0   0                            30       0    0    0    1 0 0 0             0    0
         31      0   0   0    0    0   1    0   0   0                            31       0    0    0    0 0 1 0             0    0
         32      0   0   0    0    0   0    0   0   0                            32       0    0    0    0 0 0 0             0    0
         33      0   0   0    0    0   0    0   0   0                            33       0    0    0    0 0 0 0             0    0
         34      0   0   0    0    0   0    0   0   0                            34       0    0    0    0 0 0 0             0    0
         35      0   0   0    0    0   0    0   0   0                            35       0    0    0    0 0 0 0             0    0
         36      0   0   1    0    0   1    0   0   0                            36       0    1    0    0 0 1 0             0    0
         37      0   0   0    0    0   0    0   0   0                            37       0    0    0    0 0 0 0             0    0
         38      0   0   0    0    0   0    0   0   0                            38       0    0    0    0 0 0 0             0    0
         39      0   0   0    0    0   0    0   0   0                            39       0    0    0    0 0 0 0             0    0
         40      0   0   0    0    0   0    0   0   0                            40       0    0    0    0 0 0 0             0    0
         41      0   0   0    0    0   0    0   0   0                            41       0    0    0    0 0 0 0             0    0
         42      0   0   0    0    0   0    0   0   0                            42       0    0    0    0 0 0 0             0    0
         43      0   0   0    0    0   0    0   0   0                            43       0    0    0    0 0 0 0             0    0
         44      0   0   0    0    0   0    0   0   0                            44       0    0    0    0 0 0 0             0    0
         45      0   0   0    0    0   0    0   0   0                            45       0    0    0    0 0 0 0             0    0
         46      0   0   0    0    0   0    0   0   0                            46       0    0    0    0 0 0 0             0    0
         47      0   0   0    0    0   0    0   0   0                            47       0    0    0    0 0 0 0             0    0
         48      0   0   0    0    0   0    0   0   0                            48       0    0    0    0 0 0 0             0    0
         49      0   0   0    0    0   0    0   0   0                            49       0    0    0    0 0 0 0             0    0
         50      0   0   0    0    0   0    0   0   0                            50       0    0    0    0 0 0 0             0    0
         51      0   0   0    0    0   0    0   0   0                            51       0    0    0    0 0 0 0             0    0
         52      0   0   0    0    0   0    0   0   0                            52       0    0    0    0 0 0 0             0    0

Table 14 shows the Result of the first nine patient after               Table 15 shows Result of the first nine patient after
                 training using SVM                                                    training using RBF




                                                                251                                http://sites.google.com/site/ijcsis/
                                                                                                   ISSN 1947-5500
                                                      IJCSIS) International Journal of Computer Science and Information Security,
                                                      Vol. 8, No. 5, August 2010



       MID              Patient from A to I                               VI. DISCUSSION OF FIRST FIVE PATIENTS
               A   B   C D E F G              H   I                               RESULTS WITH DOCTOR:
         1     1   0   1    0    0   1    1   1   0
         2     0   1   0    0    0   0    0   0   0                 Original Medicines given by doctor:
         3     1   1   0    1    1   1    1   1   0                      A) 1,3,5,6,14.17,19,21,23,25,26,27,29,36
         4     0   0   0    0    0   0    0   0   0                      B) 2,3,5,6,14,16,17,21,23,25,26,27,28
         5     1   1   1    1    0   1    1   1   0                      C) 1,5,6,14,25
         6     1   1   1    0    0   1    1   0   0                      D) 3,5,7,10,11,13,14,17,19,30
         7     0   0   0    1    0   0    0   0   0                      E) 3,14,15,19
         8     0   0   0    0    0   0    0   0   0                 Medicines given by the Expert system using SVM
         9     0   0   0    0    0   0    0   0   0                      A) 1,3,5,6, 16,1718,21,23,25,26,27,28,29
        10     0   0   0    1    0   0    0   0   0                      B) 1,3,5,6,16,17,18,21,23,25,26,27,28,29
        11     0   0   0    1    0   0    0   0   1                      C) 1,3,5,6,11,14,21,22,23,24,25,26,27
        12     0   0   0    0    0   0    0   0   0                      D) 1,3,5,6,13,14,17,2122,23,25,26,27,28
        13     0   1   0    1    0   0    0   0   0                      E) 1,2,3,5,14.
        14     1   1   1    1    1   0    1   0   1                 Medicines given by Expert System using RBF:
        15     0   0   0    0    1   0    0   0   0                      A) 3,5,6,14,16,17,21,23,25,26,27,28,29
        16     0   0   0    0    0   1    0   1   0                      B) 1,3,5,6,11,14,16,17,21,22,23,24,25,26,27,28,36
        17     1   1   0    1    0   1    0   1   0                      C) 1,3,5,6,14,21,25
        18     0   0   0    0    0   0    0   0   0                      D) 1,3,5,6,11,13,14,17,21,25,30
        19     1   1   0    1    1   0    0   0   0                      E) 3,5,6,14,15
        20     0   0   0    0    0   0    0   0   1                 A.        Comparative studies of SVM and RBF methods for
        21     1   1   0    0    0   1    0   0   0                 medical prescription for heart disease patient
        22     0   0   0    0    0   0    0   0   0                      RBF networks and SVM both are examples of non-linear
        23     1   1   0    0    0   0    0   0   0                 layered feed-forward networks and they are universal
        24     0   0   0    0    0   0    0   1   0                 approximations. The basic comparison of RBF and SVM NN
        25     1   1   1    0    0   1    0   1   0                 for the medical prescription for heart disease patient presented
        26     1   1   0    0    0   0    1   0   0                 in table 14 and table 15.
        27     1   1   0    0    0   0    0   1   0                      1) In both the NN model 250 data samples has given as
        28     0   0   0    0    0   0    0   0   0                           input.
        29     1   1   0    1    0   1    0   1   0                      2) The SVM model takes 250 epochs to train it while
        30     0   0   0    1    0   0    0   0   0                           RBF NN model takes only 225 epochs for the
        31     0   0   0    0    0   0    0   0   0                           training of the model.
        32     0   0   0    0    0   0    0   0   1                      3) If the training performance error is compared SVM
        33     0   0   0    0    0   0    0   0   1                           NN gives less as compared to RBF NN model.
        34     0   0   0    0    0   0    0   0   0
        35     0   0   0    0    0   0    0   0   0                      So Medicines given by the expert system using SVM is
        36     1   1   0    0    0   0    0   0   0                 not producing the appropriate result as compared with the
        37     0   0   0    0    0   0    0   0   0                 RBF.
        38     0   0   0    0    0   0    0   0   0                 B.        So we have taken doctors opinion on the result of
        39     0   0   0    0    0   0    0   0   0                 RBF as :
        40     0   0   0    0    0   0    0   0   0                 Patient A is having major heart attack for which the expert
        41     0   0   0    0    0   0    0   0   0                 system has provided the medicine no. 1 which is anxiolytic
        42     0   0   0    0    0   0    0   0   0
                                                                    and is given in almost all patients. Medicine no. 16 is
        43     0   0   0    0    0   0    0   0   0
                                                                    beneficial as it reduces the heart rate and thereby reduces
        44     0   0   0    0    0   0    0   0   0
                                                                    workload and improves outcome. Medicine no. 26 can prove
        45     0   0   0    0    0   0    0   0   0
                                                                    useful as it prevents stress erosions/ulcer. The medicine no.
        46     0   0   0    0    0   0    0   0   0
                                                                    28 is sedative (sleep inducing drug) and is beneficial and if
        47     0   0   0    0    0   0    0   0   0
                                                                    given may help improve outcome.
        48     0   0   0    0    0   0    0   0   0
                                                                    Patient B Medicine no 1 is alprazolam which is anxiolytic and
        49     0   0   0    0    0   0    0   0   0
                                                                    is given in almost all patients and won’t affect the heart
        50     0   0   0    0    0   0    0   0   0
                                                                    patient. Medicine no. 11 should not be prescribed as it is a
               0   0   0    0    0   0    0   0   0
                                                                    diuretic and can cause fall in blood pressure/electrolyte
        51
               0   0   0    0    0   0    0   0   0
                                                                    imbalance. And is not appropriate and is wrongly given by the
        52
                                                                    expert system. Medicine no 22 is Antioxidant and if given is
                                                                    useful. Medicine no. 24 is antibiotic and is given in presence
Table 16 Original medicine prescribed by the doctor.
                                                                    of infection and dose not affects the cardiac outcome.
                                                                    Patient C, Medicine 21 is prescribed which has a cardiac
                                                                    remodeling effect and if given improves outcome.


                                                              252                               http://sites.google.com/site/ijcsis/
                                                               8                                ISSN 1947-5500
                                                         IJCSIS) International Journal of Computer Science and Information Security,
                                                         Vol. 8, No. 5, August 2010



Patient D, Medicine no 1 is alprazolam which is used for its              [2]     Long, W., et.al., “Developing a program for Tracking
anxiolytic effect is given in almost all patients and wont affect                Heart Failure”, MIT Lab for Computer Science,
the heart patient. Medicine no. 6 has near to same action as                     Cambridge, MA(2001).
medicine no. 3 which is already prescribed. Expert system has             [3]     Frase, H.S.F., et.al., “Comparing complex diagnoses a
not given medicine no. 7 which is useful in this patient as                      formative evalution of the heart disease program”, MIT
patients clinical condition has poor heart pumping. Medicine                     Lab for Computer Science, Cambridge, MA(2001)
no. 19 is a antipyretic drug (to reduce fever) or as analgesic an         [4]    4. Salem, A.M. Raushdy, M and HodHod, R.A., “ A
if given wont affect the cardiac outcome. Medicine no. 25 is                     rule based expert system for diagnosis of Heart
anticlotting medicine which is wrongly given and it should be                    disease”, 8th Internationalconference on soft Computing
given in moderate to severe cases after assessing clinical                       MENDEL, Brno, Czech Rupublic, June 5-7, PP 258-
condition of the patient.                                                        263(2002).
Patient E . Medicine no. 5 is not given by the expert system              [5]     I. Turkoglu, A. Arslan, E. Ilkay, “An expert system for
which has cholesterol reducing agent and plays important role                    diagnosis of the heart valve diseases”, Expert Systems
for positive outcome. Medicine no. 6 has the same action as                      with Applications vol.23, pp. 229–236, 2002.
medicine no. 3 and is already prescribed.                                 [6]     I. Turkoglu, A. Arslan, E. Ilkay, “An intelligent system
          Medicines given by expert system in few patients are                   for diagnosis of heart valve diseases with wavelet
comparatively less and in few patients additional. The system                    packet neural networks”, Computer in Biology and
has analyzed 125 sample data and is prescribing 97% accuracy                     Medicine vol.33 pp.319–331, 2003.
in the medicines as prescribed by the doctor after his clinical           [7]    E. Comak, A. Arslan, I. Turkoglu, “A decision support
assessment. In some cases it is justified but in some cases it                   system based on support vector machines for diagnosis
depends on the Clinical condition. The additional medicines                      of the heart valve diseases”, Computers in Biology and
prescribed may prove beneficial or harmful and vice versa                        Medicine vol.37, pp.21–27, 2007.
prescribing less medicines (which if essential) can affect the            [8]    H. Uguz, A. Arslan, I. Turkoglu, “A biomedical system
cardiac outcome.                                                                 based on hidden Markov model for diagnosis of the
                       VII. CONCLUSION                                           heart valve diseases”, Pattern Recognition Letters
In this paper, around 300 patient’s information is collected                     vol.28 pp.395–404, 2007.
from Sahara Hospital, under supervision of Dr. Abdul Jabbar,              [9]    A. Sengur, “An expert system based on principal
(MD Medicine) Sahara Hospital, Roshan Gate, Aurangabad.                          component analysis, artificial immune system and
The collected information is coded, normalized and entered                       fuzzy k-NN for diagnosis of valvular heart diseases”,
into 13 different excel sub-sheets. All the patients data is                     Computers in Biology and Medicine vol.38 pp.329–
trained by using SVM and RBF. Around 50 samples were                             338, .2008.
tested with these two techniques. If the more data set is used            [10]    A. Sengur, “An expert system based on linear
for the training the NN model gives more robust results. The                     discriminant analysis and adaptive neuro-fuzzy
analysis model by using SVM and RBF of ANN gives better                          inference system to diagnosis heart valve diseases”,
result for medical prescription for heart disease patient.                       Expert Systems with Applications vol.35 pp.214–222,
However, there are several techniques that can improve the                       2008.
speed and performance of the back propagation algorithm,                  [11]    P.I.J. Keeton, F.S. Schlindwein, “Application of
weight initialization, use of momentum and adaptive learning                     Wavelets in Doppler Ultrasound, vol. 17, number 1,
rate. It is found that the result of testing data by using SVM is                MCB University Press, pp. 38–45, 1997.
not satisfactory but the medicines prescribed by the RBF are              [12]    I.A. Wright, N.A.J. Gough, F. Rakebrandt, M. Wahab,
satisfactory as per the result verified by the doctor. The                       J.P. Woodcock, “Neural network analysis of Doppler
diagnosis performances of this study shows the advantage of                      ultrasound blood flow signals: a pilot study, Ultrasound
this system. : it is rapid, easy to operate, non-invasive and not                in Medicine & Biology vol.23,pp. 683–690, 1997.
expensive. The working prototype model in the field of heart              [13]    B. C. B chan ,F.H.Y. Chan ,F.K. Lam
diagnosis can use the system. It also helps for training                         ,P.W.Lui,P.W.F.Poon, “”Fast detection of venous air
begineer’s doctors and medical students who work in the field                    embolism is Doppler heart sound using the wavelet
of heart diagnosis. In future, this work may be extend using                     transform”, IEEE Transaction on               Biomedical
regression technique.                                                            Engineering vol. 44 pp.237-245, (1997).
                   ACKNOWLEDGEMENT                                        [14]    I. Guler, M.K. Kiymik, S. Kara, M.E. Yuksel,
          The author thank to Dr. Abdul Jabbar (MD                               “Application of autoregressive analysis to 20 MHz
Medicine), Sahara Hospital, Raushan Gate, Aurangabad for                         pulsed Doppler data in real time, International Journal
supervision of medicine data as well as evaluation of medicine                   Biomedical Computing vol.31,pp.247–256, 1992.
provided by the expert system.                                            [15]   http://elearning.najah.edu/OldData/pdfs/Intro%20Exper
                         REFERENCES                                              t%       20     Systems%20test-me.co.uk.       ppt#262,5,
[1]    Fraser, H.S.F., et al., “Differential diagnoses of the                    Components of an Expert System.
       heart disease program have better senditivity than                 [16]    J. Galindo, P. Tamayo, Credit risk assessment using
       Resident Physicians”, Tufts-New England Medical                           statistical and machine learning: basic methodology and
       Center, Boston, MA(2001).                                                 risk modeling applications, Computational Economics
                                                                                 15 (1 – 2) (2000) 107–143.


                                                                    253                               http://sites.google.com/site/ijcsis/
                                                                                                      ISSN 1947-5500
                                                       IJCSIS) International Journal of Computer Science and Information Security,
                                                       Vol. 8, No. 5, August 2010



[17]   D. Michie, D.J. Spiegelhalter, C.C. Taylor, Machine [25] E. Osuna, R. Freund, F. Girosi, Training support vector
      Learning, Neural and Statistical Classification, Elis             machines: an application to face detection, Proceedings
      Horwood, London, 1994.                                            of Computer Vision and Pattern Recognition, 1997, pp.
[18] S.M. Weiss, C.A. Kulikowski, Computer Systems That                 130–136.
      Learn: Classification and Prediction Methods from [26] F.E.H. Tay, L.J. Cao, Modified support vector
      Statistics, Neural Networks, Machine Learning and                 machines in financial time series forecasting,
      Expert Systems, Morgan Kaufmann, San Mateo, 1991.                 Neurocomputing 48 (2002) 847– 861.
[19] V. Vapnik, The Nature of Statistical Learning Theory, [27] T. Van Gestel, J.A.K. Suykens, D.-E. Baestaens, A.
      Springer- Verlag, New York, 1995.                                 Lambrechts, G. Lanckriet, B. Vandaele, B. De Moor, J.
[20] M.A. Hearst, S.T. Dumais, E. Osman, J. Platt, B.                   Vandewalle, Financial time series prediction using least
      Scho¨lkopf, Support vector machines, IEEE Intelligent             squares support vector machines within the evidence
      Systems 13 (4) (1998) 18– 28.                                     framework, IEEE Transactions on Neural Networks 12
[21] M.P. Brown, W.N. Grudy, D. Lin, N. Cristianini, C.W.               (4) (2001) 809– 821.
      Sugnet, T.S. Furey, M. Ares, D. Haussler, Knowledge- [28] N. Cristianini, J. Shawe-Taylor, An Introduction to
      based analysis of microarray gene expression data by              Support Vector Machines, Cambridge Univ. Press,
      using support vector machines. Proceedings of National            Cambridge, New York, 2000.
      Academy of Sciences 97 (1) (2000) 262–267.                 [29] Broomhead D., & Lowe, D., Multivariable functional
[22] T.S. Jaakkola, D. Haussler, Exploiting generative                  interpolation and adaptive networks. Complex Systems,
      models in discriminative classifiers, in: M.S. Kearns,            vol.2, pp.321-355, 1988.
      S.A. Solla, D.A. Cohn (Eds.), Advances in Neural [30] Haralambos Sarimveis, Philip Doganis, Alex
      Information Processing Systems, MIT Press,                        Alexandridis, “A classification technique based on
      Cambridge, 1998.                                                  radial basis function neural networks”, Advances in
[23] A. Zien, G. Ra¨tsch, S. Mika, B. Scho¨lkopf, T.                    Engineering Software vol.37, pp.218–221, 2006.
      Lengauer, K.-R. Mu¨ller, Engineering support vector
      machine kernels that recognize translation initiation
      sites, Bioinformatics 16 (9) (2000) 799– 807.
[24] T. Joachims, Text categorization with support vector
      machines, Proceedings of the European Conference on
      Machine Learning (ECML), 1998.
                                                     AUTHORS PROFILE
Shaikh Abdul Hannan (Aurangabad, 02-05-1973) is a Member of IETE, ISTE. He has received the B.Sc. and M.Sc. degree
                 with first class from Dr. Babasaheb Ambedkar Marathwada University, Aurangabad. He has completed M.Phil
                 with first class from Y.C.M.O, Nasik and persuing Ph.D. from North Maharashtra University, Jalgoan. He is
                 lecturer in computer department of Vivekanand College, Aurangabad since last 10 years. He has published more
                 than 40 papers in various international conferences and journals. His has taken Ist prize in International
                 conference at Nagpur (ITBI-2009). He is board member of “Journal of Advanced Research in Computer
                 Engineering: An International Journal” New Delhi, India.


                Dr. V. D. Bhagile received the M.C.M degree in 2003 & Ph.D. from Dr. Babasaheb Ambedkar Marathwada
                University, Aurangabad, Maharashtra, INDIA, She is recipient of Award of Golden Jubilee University Junior
                Research fellowship for the years 2007-08 & 2008-09. Presently working as a lecturer & InCharge Principal in
                Sayali Charitable Trust's College of Computer Science and Information Technology (612), Aurangabad.



               Dr. R. R. Manza received M.Sc. degree in 1998 from Dr. Babasaheb Ambedkar Marathwada University,
               Aurangabad, Maharashtra India. He has passed NET/SET in 2002 from Pune university and UGC, New Delhi.
               He has completed one minor research project on “Development of Biomedical Monitoring System for Diabetes
               using Nanorobotics in 2008-2009. He has published more and 80 papers in International Journals and
               conferences. Presently working as a Asst. Professor in Department of Computer Science and I.T., Dr. B.A.M.
               University, Aurangabad.

                Dr. R. J Ramteke received M.Sc. degree in 1995 & Ph.D. from Dr. Babasaheb Ambedkar Marathwada
                University, Aurangabad, Maharashtra, India. He has taken awards from Teacher Research Fellowship under
                Faculty Improvement Programme Xth Plan of University Grants Commission, New Delhi during July 2004-
                July2006 and Finalist for Best Student Paper Award IEEE International Conference on Cybernetics and
                Intelligent System (CIS-2006), Bangkok, Thailand, on 7th – 9th June 2006. He has published more and 65
                papers in International Journals and conferences. Presently working as Reader in Department of Computer
                Science, North Maharashtra University, Jalgaon.

                                                               254                               http://sites.google.com/site/ijcsis/
                                                                                                 ISSN 1947-5500

				
DOCUMENT INFO
Description: Vol. 8 No. 5 August 2010 International Journal of Computer Science and Information Security Publication August 2010, Volume 8 No. 5 (Download Full Journal) (Archive)