Journal of Software Engineering and Applications, 2013, 6, 85-97
http://dx.doi.org/10.4236/jsea.2013.63013 Published Online March 2013 (http://www.scirp.org/journal/jsea)
85
Comparison of Various Classification Techniques Using
Different Data Mining Tools for Diabetes Diagnosis
Rashedur M. Rahman, Farhana Afroz
Department of Electrical Engineering and Computer Science, North South University, Dhaka, Bangladesh.
Email: rashedur@northsouth.edu, farhana_sl_03@yahoo.com
Received December 21st, 2012; revised January 20th, 2013; accepted January 30th, 2013
ABSTRACT
In the absence of medical diagnosis evidences, it is difficult for the experts to opine about the grade of disease with af-
firmation. Generally many tests are done that involve clustering or classification of large scale data. However many
tests could complicate the main diagnosis process and lead to the difficulty in obtaining the end results, particularly in
the case where many tests are performed. This kind of difficulty could be resolved with the aid of machine learning
techniques. In this research, we present a comparative study of different classification techniques using three data min-
ing tools named WEKA, TANAGRA and MATLAB. The aim of this paper is to analyze the performance of different
classification techniques for a set of large data. A fundamental review on the selected techniques is presented for intro-
duction purpose. The diabetes data with a total instance of 768 and 9 attributes (8 for input and 1 for output) will be
used to test and justify the differences between the classification methods. Subsequently, the classification technique
that has the potential to significantly improve the common or conventional methods will be suggested for use in large
scale data, bioinformatics or other general applications.
Keywords: Classification; Neural Network; Fuzzy Logic; Decision Tree; Performance Measurement
1. Introduction
The aim of this study is to investigate the performance of
different classification methods using WEKA, TANA-
GRA and MATLAB tool on Diabetes Dataset, specifi-
cally Pima Indian Diabetes Dataset. A major problem in
bioinformatics analysis or medical science is to attain the
correct diagnosis of certain important information. For
the ultimate diagnosis, generally many tests are done that
involve clustering or classification of large scale data. All
of these test procedures are said to be necessary in order
to reach the ultimate diagnosis. However, on the other
hand, too many tests could complicate the main diagnosis
process and lead to the difficulty in obtaining the end
results, particularly in the case where many tests are per-
formed. This kind of difficulty could be resolved with the
aid of machine learning which could be used directly to
obtain the end result with the aid of several artificial in-
telligence techniques. Machine learning covers such a
broad range of processes that it is difficult to define it
precisely. A dictionary definition includes phrases such
as to gain knowledge or understanding of or skill by
studying the instruction or experience and modification
of a behavioral tendency by experienced zoologists and
psychologists study learning in animals and humans [1].
The extraction of important information from a large pile
of data and its correlations is often the advantage of us-
ing machine learning. New knowledge about tasks is
constantly being discovered by humans. There is a con-
stant stream of new events in the world and continuing
redesign of Artificial Intelligent systems to conform to
new knowledge is impractical but machine learning
methods might be able to track much of it [1].
There is a substantial amount of research has been
done with machine learning algorithms such as Bayes
network, Multilayer Perceptron, Decision tree and prun-
ing like J48graft, C4.5, Single Conjunctive Rule Learner
like FLR, JRip and Fuzzy Inference System and Adap-
tive Neuro-Fuzzy Inference System.
2. Related Work
A good number of researches have been reported in lit-
erature on diagnosis of different deceases. Sapna and
Tamilarasi [2] proposed a technique based on neuropathy
diabetics. Nerve disorder is caused by diabetic mellitus.
Long term diabetic patients could have diabetic neu-
ropathies very easily. There is fifty (50%) percent prob-
ability to have such diseases which affect many nerves
system of the body. For example, body wall, limbs
(which is called as somatic nerves) could be affected. On
the other hand, internal organ like heart, stomach, etc.,
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
86
are known as automatic nerves. In this paper the risk
factors and symptoms of diabetic neuropathy are used to
make the fuzzy relation equation. Fuzzy relation equation
is linked with the perception of composition of binary
relations that means they used Multilayer Perceptron NN
using Fuzzy Inference System.
Leonarda and Antonio [3] proposed automatic detec-
tion of diabetic symptoms in retinal images by using a
multilevel perceptron neural network. The network is
trained using algorithms for evaluating the optimal global
threshold which can minimize pixel classification errors.
System performances are evaluated by means of an ade-
quate index to provide percentage measure in the detec-
tion of eye suspect regions based on neuro-fuzzy subsys-
tem.
Radha and Rajagopalan [4] introduced an application
of fuzzy logic to diagnosis of diabetes. It describes the
fuzzy sets and linguistic variables that contribute to the
diagnosis of disease particularly diabetes. As we all
know fuzzy logic is a computational paradigm that pro-
vides a tool based on mathematics which deals with un-
certainty. At the same time this paper also presents a
computer-based Fuzzy Logic with maximum and mini-
mum relationship, membership values consisting of the
components, specifying fuzzy set frame work. Forty pa-
tients’ data have been collected to make this relationship
more strong.
Jeatrakul and Wong [5] presented a comparison of
neural network techniques for binary classification prob-
lems. The classification performance obtained by five
different types of neural networks, i.e., Back Propagation
Neural Network (BPNN), Radial Basis Function Neural
Network (RBFNN), General Regression Neural Network
(GRNN), Probabilistic Neural Network (PNN), and
Complementary Neural Network (CMTNN). The com-
parison is done based on three benchmark data sets ob-
tained from UCI machine learning repository.
Zhou, Purvis and Kasabov [6] described a general
method based on the statistical analysis of training data
to select fuzzy membership functions to be used in con-
nection with fuzzy neural networks. The technique is first
described and then illustrated by means of two experi-
mental examinations for medical data.
Ten-Ho Lin and Von-Wun Soo [7] proposed alterna-
tive pruning method based on the Minimal Description
Length (MDL) principle. The MDL principle can be
viewed as a tradeoff between theory complexity and data
prediction accuracy given the theory. A greedy search
algorithm of the minimum description length to prune the
fuzzy ARTMAP categories one by one was proposed.
The experiments showed that fuzzy ARTMAP pruned
with the MDL principle gave better performance with
fewer categories created compared to original fuzzy
ARTMAP and other machine learning systems. They
tested those techniques on a number of benchmark clini-
cal databases such as heart disease, breast cancer and
diabetes databases.
Faezeh, Hossien, Ebrahim [8] proposed a fuzzy clus-
tering technique (FACT) which determined the number
of appropriate clusters based on the pattern essence. Dif-
ferent experiments for algorithm evaluation were per-
formed which showed a better performance compared to
the typical widely used K-means clustering algorithm.
Data was taken from the UCI Machine Learning Reposi-
tory [9].
Santi Waulan et al. [10] proposed a new SSVM for
classification problems. It is called Multiple Knot Spline
SSVM (MKS-SSVM). To evaluate the effectiveness of
their method, they carried out an experiment on Pima
Indian diabetes dataset. First, theoretical of MKS-SSVM
was presented. Then, application of MKS-SSVM and
comparison with SSVM in diabetes disease diagnosis
was given. The results of this study showed that the
MKS-SSVM was effective to detect diabetes disease
diagnosis.
3. Data Set Description
The characteristics of the data set used in this research
are summarized in following Table 1. The detailed de-
scriptions of the data set are available at UCI repository
[9].
The objective of this data set was diagnosis of diabetes
of Pima Indians. Based on personal data, such as age,
number of times pregnant, and the results of medical
examinations, e.g., blood pressure, body mass index,
result of glucose tolerance test, etc., it is tried to decide
whether a Pima Indian individual was diabetes positive
or not.
Pima Indian Diabetes Data (PIDD) set is publicly
available from the machine learning database at UCI. All
patients represented in this data set are females with at
least 21 years old of Pima Indian heritage living near
Phoenix, Arizona. The problem posed here is to predict
whether a person would test positive given a number of
physiological measurements and medical test results.
This is a two-class problem with class value 1 being in-
terpreted as “tested positive for diabetes”. There are 500
Table 1. Characteristics of data sets.
Data set No. of example Input attributes Output classesTotal No. of
attributes Missing attributes status Noisy attributes
status
Pima 768 8 2 9 No No
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis 87
examples of class 1 and 268 of class 2.
This data set is extracted from a larger database origin-
nally owned by the National Institute of Diabetes and
Digestive and Kidney Diseases. The purpose of the study
is to investigate the relationship between the diabetes
diagnostic result and a list of variables that represent
physiological measurements and medical attributes. The
data set in the UCI repository contains 768 observations
and 9 variables with no missing values reported. How-
ever, as some researchers point out, there are a number of
impossible values, such as 0 body mass index and 0
plasma glucose. Furthermore, one attribute (2-hour se-
rum insulin) contains almost 50% impossible values. To
keep the sample size reasonably large, this attribute is
removed from analysis. There are 236 observations that
have at least one impossible value of glucose, blood
pressure, triceps skin thickness, and body mass index.
There are nine variables, including the binary response
variable, in this data set; all other attributes are numeric-
valued. The attributes are given below:
1) Number of times pregnant
2) Plasma glucose concentration a 2 hours in an oral
glucose tolerance test
3) Diastolic blood pressure (mm Hg)
4) Triceps skin fold thickness (mm)
5) 2-hour serum insulin (mu U/ml)
6) Body mass index (weight in kg/(height in m)^2)
7) Diabetes pedigree function
8) Age (years)
9) Class variable (0 or 1)
4. Methodology
We use different classification techniques in this research.
Those techniques with running parameters are given
below:
4.1. Multilayer Perceptron
Multilayer perceptron (MLP) [11] is one of the most
commonly used neural network classification algorithms.
The architecture used for the MLP during simulations
with PIDD dataset consisted of a three layer feed-for-
ward neural network: one input, one hidden, and one
output layer. Selected parameters for the model are:
learningRate = 0.3/0.15; momentum = 0.2; randomSeed
= 0; validationThreshold = 20, Number of Epochs = 500.
4.2. BayesNet
BayesNet [12] learns Bayesian networks under the pre-
sumptions: nominal attributes (numeric one are pre-de-
scretized) and no missing values (any such values are
replaced globally). There are two different parts for es-
timating the conditional probability tables of the network.
In this study we run BayesNet with the SimpleEstimator
and K2 search algorithm without using ADTree. K2 al-
gorithm is a greedy search algorithm that works as fol-
lows. Suppose we know the total ordering of the nodes.
Initially each node has no parents. The algorithm then
incrementally adds the parent whose addition increases
most of the score of the resulting structure. When no ad-
dition of a single parent can increase the score, it stops
adding parents to the node. Since an ordering of the
nodes is known beforehand, the search space under this
constraint is much smaller than the entire space. And we
do not need to check for cycles, since the total ordering
guarantees that there is no cycle in the deduced structures.
Furthermore, based on some appropriate assumptions, we
can choose the parents for each node independently.
4.3. Naïve Byes
The Naïve Bayes [12] classifier provides a simple ap-
proach, with clear semantics, representing and learning
probabilistic knowledge. It is termed naïve because is
relies on two important simplifying assumes that the pre-
dictive attributes are conditionally independent given the
class, and it assumes that no hidden or latent attributes
influence the prediction process.
4.4. J48graft (C4.5 Decision Tree Revision 8)
Perhaps C4.5 algorithm which was developed by Quinlan
[13] is the most popular tree classifier till today. Weka
classifier package has its own version of C4.5 known as
J48 or J48graft. For this study, C4.5 classifier used in
TANAGRA platform and J48graft classifier used in
WEKA platform. J48graft is an optimized implemen-
tation of C4.5 rev. 8. J48graft is experimented is this
study with the parameters: confidenceFactor = 0.25;
minNumObj = 2; subtreeRaising = True; unpruned =
False. C4.5 is experimented in this study with the pa-
rameters: Min size of leaves = 5; Confidence-Level for
pessimistic = 0.25. Final decision tree built from the al-
gorithm is depicted in Figure 1.
4.5. Fuzzy Lattice Reasoning (FLR)
The Fuzzy Lattice Reasoning (FLR) classifier is pre-
sented for inducing descriptive, decision-making knowl-
edge (rules) in a mathematical lattice data domain in-
cluding space RN. Tunable generalization is possible
based on non-linear (sigmoid) positive valuation func-
tions; moreover, the FLR classifier can deal with missing
data. Learning is carried out both incrementally and fast
by computing disjunctions of join-lattice interval con-
junctions, where a join-lattice interval conjunction cor-
responds to a hyperbox in RN. In this study we evaluated
FLR classifier in WEKA with the parameters: Rhoa = 0.5;
Number of Rules = 2.
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
88
Figure 1. Decision tree for J48graft algorithm for PIDD after experiment with WEKA.
4.6. JRip (JRipper)
Repeated Incremental Pruning to Produce Error Reduc-
tion (RIPPER) [14] is one of the basic and most popular
algorithms. Classes are examined in increasing size and
an initial set of rules for the class is generated using in-
cremental reduced-error pruning. In this study, we evalu-
ated RIPPER through JRip, an implementation of RIP-
PER in WEKA with the parameters: folds = 10; minNo =
2; optimizations = 2; seed = 1; usePruning = true.
4.7. Fuzzy Inference System (FIS)
Fuzzy inference system (FIS) is a technology developed
for granular rule induction and generalization based on
fuzzy logic. Note that since a data cluster can be interp-
reted as a (fuzzy) granule, data clustering may be closely
related to fuzzy rule induction. Neural implementations
have provided conventional FIS a capacity for parallel
implementation.
4.8. Adaptive Neuro-Fuzzy Inference System
(ANFIS)
In this work uses ANFIS (Adaptive Neuro-Fuzzy Infer-
ence System), a fuzzy classifier that is part of the
MATLAB Fuzzy Logic Toolbox [15]. ANFIS (Adaptive-
Network-Based Fuzzy Inference System) is a fuzzy in-
ference system implemented under the framework of
adaptive networks [16]. ANFIS is a type of Neuro-Fuzzy
network which has the fuzzy rules embedded within the
neural network.
5. Performance Metrics
We measure the performance of the classifiers with
respect to different performance metrics like accuracy,
precision, recall, F-measure, area under ROC curve, and
gamma statistic. More details of those measures could be
found elsewhere [14,17,18].
The mean absolute error Ei of an individual program i
is evaluated by the Equation (1):

1
1n
i
ij
j
EP
nj
T
(1)
where P(ij) is the value predicted by the individual
program i for sample case j (out of n sample cases); and
Tj is the target value for sample case j. For a perfect fit,
P(ij) = Tj and Ei = 0. So, the Ei index ranges from 0 to
infinity, with 0 corresponding to the ideal.
The relative absolute error is also relative to a simple
predictor, which is just the average of the actual values.
In this case, though, the error is just the total absolute
error instead of the total squared error. Thus, the relative
absolute error takes the total absolute error and norma-
lizes it by dividing by the total absolute error of the
simple predictor.
Mathematically, the relative absolute error Ei of an
individual program i is evaluated by the Equation (2):

11
nn
ij
ij
jj
EPTT

j
T

(2)
where P(ij) is the value predicted by the individual pro-
gram i for sample case j (out of n sample cases); Tj is the
target value for sample case j; andTis given by the for-
mula (Equation (3)):
1
1n
j
j
T
n
T
(3)
We also calculate the Root Mean Squared Error which
is the square root of Equation (3) and Root Relative
Squared Error is square root of Equation (4).
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis 89
6. Survey Result
To measure and investigate the performance on the se-
lected classification methods namely Multilayer Percep-
tron (MLP) Neural Network, Bayes Network Classifier,
J48graft (C4.5 Decision Tree Revision 8), JRip (RIP-
PER), Fuzzy Lattice Reasoning (FLR) classifier, we use
the experiment procedures by WEKA whereas MLP,
Naïve Bayes, C4.5 Decision Tree provided by TANA-
GRA and lastly Fuzzy Inference System (FIS), Adaptive
Neuro-Fuzzy Inference System (ANFIS) experiment
procedure provided by MATLAB. The 66% data is used
for training and the remaining is for testing purposes that
is shown in Table 2.
In this study, all data is considered as instances and
features in the data are known as attributes. The simula-
tion results are partitioned into several sub items for eas-
ier analysis and evaluation. Different performance matrix
like TP rate, FP rate, and Precision, Recall, F-measure
and ROC area are presented in numeric value during
training and testing phase. The summary of those results
by running the techniques in WEKA is reported in Table
3. Then we run the algorithms in Tanagra and results are
reported through Tables 4 and 5. Performance measure-
ment in Matlab environment is reported in Table 6. Ta-
bles 7 and 8 report different types of error measurement
like mean absolute error and root mean squared error, the
time taken to build model in second and Kappa statistic.
Table 9 reports accurate and error rate that is repre-
sented in percentage value. Finally, Table 10 shows the
rules that used FIS and ANFIS for MATLAB.
Table 2. Number of instances in the training and test data
set.
Data set No. of training
data No. of test
data Total
Pima Indians
Diabetes 507 261 768
Table 3. Different performrance me tr ic s r unning in WEKA.
Classifier Phase TP rate FP rate Precision Recall F-measure ROC Area
Train 0.806 0.191 0.819 0.806 0.809 0.872
MLP
Test 0.778 0.306 0.774 0.778 0.776 0.813
Train 0.783 0.26 0.783 0.783 0.783 0.851
BayesNet
Test 0.797 0.253 0.799 0.797 0.798 0.848
Train 0.841 0.241 0.842 0.841 0.836 0.888
J48graft
Test 0.785 0.189 0.816 0.785 0.792 0.803
Train 0.794 0.257 0.792 0.794 0.793 0.785
JRip
Test 0.824 0.294 0.821 0.824 0.816 0.766
Train 0.358 0.344 0.774 0.358 0.2 0.507
FLR
Test 0.67 0.662 0.582 0.67 0.572 0.504
Table 4. Different performance metrics in TANAGRA.
Classifier Recall Precision
MLP 0.8275 0.8275
Naïve bayes 1 1
C4.5 0.90465 0.90465
Table 5. Performance measuring in training and test data
set using TANAGRA.
Classifier Accuracy Error rate Time (seconds)
MLP 83.85% 16.15% 2.36
Naïve bayes100.00% 0.00% 0.001
C4.5 90.63% 9.38% 0.031
Table 6. Performance measur ing in r ule base d fuzzy appr oac h using MATLAB.
Learning systems Training/test epochs Avg. error after
training/test No. of
extracted rules Rules ac cur acy (%)
Fuzzy Inference System 500 7.6358 7 71.51
Adaptive Neuro-Fuzzy
Inference System 500 7.6358 7 78.79
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
90
Table 7. Error measurement for different classifiers in WEKA.
Classifier Phase M ean absolute
error Avg. MAE Root mean
squared errorAvg.
RMSE Relative absolute
error Avg. RAE Root relative
squared error Avg. RrSE
Training 0.2852 0.3815 62.75% 80.05%
MLP
Testing 0.2892
0.2872
0.4007
0.3911
64.08%
63.42%
85.44%
82.75%
Training 0.2707 0.3878 59.55% 81.36%
BayesNet
Testing 0.2816
0.2762
0.3867
0.3873
62.40%
60.98%
82.44%
81.9%
Training 0.2383 0.3452 52.43% 72.42%
J48graft
Testing 0.2914
0.2649
0.4013
0.3733
64.57%
58.50%
85.57%
79%
Training 0.3091 0.3931 68.02% 82.48%
JRip
Testing 0.3165
0.3128
0.3842
0.3887
70.17%
69.1%
82.07%
82.28%
Training 0.6419 0.8012 141.24% 168.09%
FLR
Testing 0.3295
0.4857
0.5740
0.6876
73.06%
107.15%
122.63%
145.36%
Table 8. Performance measuring in training and test data set using WEKA.
Classifier Phase AccuracyAvg. AC Error RateAvg. ERTime (seconds) Avg. TT Kappa st atistic Avg. KS
Training 80.60% 19.40% 101.08 0.5904
MLP
Testing 77.78%
79.19%
22.22%
20.81%
25.17
63.13
0.4812
0.5358
Training 78.26% 21.74% 0.03 0.5218
BayesNet
Testing 79.69%
78.98%
20.31%
21.03%
0.05
0.04
0.5391
0.5305
Training 84.11% 15.89% 0.19 0.6319
J48graft
Testing 78.54%
81.33%
21.46%
18.68%
0.08
0.135
0.5481
0.59
Training 79.43% 20.57% 0.3 0.5425
JRip
Testing 82.38%
80.91%
17.62%
19.10%
0.28
0.29
0.5658
0.5542
Training 35.81% 64.19% 0.03 0.0098
FLR
Testing 67.05%
51.43%
32.95%
48.57%
0.02
0.025
0.0115
0.0107
Table 9. Accuracy in percentage for different classification comparing three tools.
Tool MLP BayesNet/Naïve bayes C4.5/J48graftJRip FLR FIS ANFIS Average
WEKA 79.19% 78.98% 81.33% 80.91% 51.43% nill nill 74.37%
TANAGRA 83.85% 100% 90.63% nill nill nill nill 91.49%
MATLAB nill nill nill nill nill 71.51% 78.79% 75.15%
Table 10. Sample rules framed for the proposed FIS and ANFIS.
IF THEN
Rule No. preg. plas bp skin insl bmi dpf age class 0 (weight) class 1 (weight)
1 0 103 >40 26 156 35.3 0.179 34 0.955 0.5
2 3 not define not define 35 >156 35.3 0.787 not define0.5 0.928
3 not define not define not define not definenot definenot define0.179 34 0.955 0.5
4 not define 103 not define not definenot definenot define0.787 not define0.944 0.5
5 not define not define not define not define156 35.3 not define>34 or 370.912 0.5
6 not define >135 not define not define185 >33.7 1.096 >37 0.5 0.928
7 6 >103 not define not definenot define>35.3 1.096 >34 0.5 0.909
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
Copyright © 2013 SciRes. JSEA
91
7. Result Analysis and Discussion
In this study, we examine the performance of different
classification methods that could generate accuracy and
some error to diagnosis the data set.
According to Figu r e 2 and Table 8, we can clearly see
the highest accuracy is 81.33% belongs to J48graft and
lowest accuracy is 51.43% that belongs to FLR. The total
time required to build the model is also a crucial pa-
rameter in comparing the classification algorithm.
Based on Figure 3 and Table 7, we can compare er-
rors among different classifiers in WEKA. We clearly
find out that J48graft is the best, second best is the Bayes
Net and MLP & JRip is moderate but FLR is arguable.
An algorithm which has a lower error rate will be pre-
ferred as it has more powerful classification capability
and ability in terms of medical and bioinformatics fields.
Figure 4. Time comparison of WEKA & TANAGRA
gorithm is around 0.01 - 0.59. Based on the Kappa Statis-
tic criteria, the accuracy of this classification purposes is
substantial [14]. So according to best average kappa sta-
tistic the J48graft classifier is best among others. In Ta-
ble 5 investigated the accuracy using TANAGRA tools
for three classifiers like MLP, Naïve Bayes and C4.5.
According to the TANAGRA and algorithms, Naïve
Bayes is best comparatively others classifiers cause
100% accuracy achieved by Naïve Bayes and take time
to build model is 0.001 s that also lowest time compare
to others.
From Figure 4, we see that FLR classifier requires the
shortest time which is around 0.025 seconds compared to
other. MLP algorithm requires the longest model build-
ing time which is around 63.13 seconds. The second one
the list is Bayes network with 0.04 seconds, whereas
J48graft takes 0.135 seconds.
Kappa statistic is used to assess the accuracy of any
particular measuring cases, it is usual to distinguish be-
tween the reliability of the data collected and their valid-
ity [18]. The average Kappa score from the selected al-
Cluster analysis is a way to examine similarities and
dissimilarities of observations or objects. Data often fall
naturally into groups, or clusters, of observations, where
the characteristics of objects in the same cluster are
similar and the characteristics of objects in different
clusters are dissimilar. In this examination, cluster analy-
sis is used for the purpose of segregating the patients
with high risk and low risk. Grouping of clusters are used
to identify the patients who need the emergency care.
Using the MLP, BayesNet, JRip, FLR or J48graft in
WEKA using EM cluster are shaped by training them
with input/output data rather than specifying them auto-
matically. Figure 5 shows the WEKA cluster analysis.
Another figure generated from MATLAB result is
shown in Table 7. Since same iteration and same dataset
used, average error is similar. But rules accuracy is
71.51% and 78.79% of FIS and ANFIS respectively for
different network architecture. Fuzzy Inference System
(FIS) and Adaptive Neuro-Fuzzy Inference Systems
(ANFIS) classifiers are chosen from MATLAB fuzzy
logic toolbox. IF-THEN rules are used in adaptive classi-
fiers. For performance measurement, we use 7 IF-THEN
fuzzy rules and mamdani membership for FIS and
sugeno for ANFIS. Figure 6 shows the FIS model with
mamdani membership function.
Figure 2. Accuracy of three tools using PIDD.
Figure 7 shows the membership function for input and
output variable. Eight attributes need different member-
ship function according to rules that showed in Table 10.
All membership functions are continuous, all member-
ship functions map an interval [a, b] to [0, 1], μ [a, b] to
Figure 3. Error comparison for WEKA.
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
92
Figure 5. Generation of cluster for WEKA using PIDD.
Figure 6. Fuzzy Inference System.
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis 93
Figure 7. Membership func tion of input and output.
[0, 1]. Figure 7 illustrates plasma attribute that models
gbellmf function (Generalized bell-shaped-in member-
ship function) for condition. The generalized bell func-
tion depends on three parameters a, b, and c as given by

2
1
;,,
1
b
fxabc
x
c
a
where the parameter b is usually positive. The parameter
c locates the center of the curve. We have to call the
function “gbellmf” with the parameter vector params
whose entries are a, b, and c, respectively. The syntax of
gbellmf is y = gbellmf (x, params).
Example:

positive 2
1,if 103
1,if 103
103
1
x
x
xx
a




negative 2
0,if 135
1, if135
135
1
x
x
xx
a



The second part of Figure 7 illustrates the class 0
output of plasma input, membership function according
to diabetes condition or rules. For class output four mem-
bership functions are used. For example, zmf (Z-shaped
built-in membership function), gaussmf (Gaussian curve
built-in membership function), gbellmf (Generalized
bell-shaped-in membership function) and trimf (Trian-
gular-shaped built-in membership function) are used. 0,
0.912, 0.955 and 0.944 plasma output for zmf, gaussmf,
gbellmf and trimf membership function respectively.
The rule base consists of a set of fuzzy propositions
and is derived from the knowledge base of the medical
expertise. A fuzzy proposition or a statement establishes
a relationship between different input fuzzy sets and out-
put fuzzy sets. In this phase, the decision rules are con-
structed for input parameter and control output values to
find the positive or negative diabetes. In order to validate
the fuzzy logic approach used in construction of Fuzzy
Inference System, the extensive simulation is carried out
on the simulated model.
The system responses with variations defined in the
membership functions as a rule viewer, surface viewer.
Data training, checking, testing with sample data is done
to capture the error by ANFIS using FIS model. ANFIS
uses hybrid learning rules, which combines the gradient
method and the least squares estimate to identify pa-
rameters. The ANFIS network structure used in this re-
search is depicted in Figure 8.
The rule-base constructed are simulated using MAT-
LAB to identify the output parameter class 0 or class 1
that means tested positive and negative. Figures 9 and 10
finally report the simulation view of Fuzzy Inference
System and Adaptive Neuro-Fuzzy Inference System
(ANFIS). In ANFIS Editor, membership functions are
shaped by training them with input/output data rather
than specifying them manually. The training algorithm is
a combination (hybrid) with a least square system to
learn from the data.
To measure the diagnosis accuracy with Fuzzy ap-
proach, training and testing iteration is 500 epochs (for
all tools and all classifiers). By ANFIS we achieve
78.79% accuracy that’s highest than FIS (71.51%). Ta-
ble 10 reports the diabetes diagnosis rules.
Finally, Figures 11 and 12 show the surface construc-
tion of negative and positive class.
8. Conclusion and Future Work
The objective of this study is to evaluate and investigate
nine selected classification algorithms based on WEKA,
TANAGRA and MATLAB. The best algorithm in
WEKA is J48graft classifier with an accuracy of 81.33%
that takes 0.135 seconds for training. In TANAGRA,
Naïve Bayes classifier provides accuracy of 100% with
training time 0.001 seconds. In MATLAB, ANFIS has
78.79% accuracy If we compare with average accuracy,
TANAGRA machine learning tool is the best compared
to WEKA and MATLAB. Those results suggest that
among the machine learning algorithm tested on PIDD,
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
94
Figure 8. ANFIS network structure.
Figure 9. Simulation view of rules base of diabetes using FIS.
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis 95
Figure 10. Simulation view of rule base of diabetes using ANFIS.
Figure 11. Surface construc tion of F IS and ANF IS with skin vs. bmi.
Copyright © 2013 SciRes. JSEA
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis
Copyright © 2013 SciRes. JSEA
96
Figure 12. Surface construc tion of F IS and ANF IS with plas vs. ins.
Bayes network classifier like NaiveBayes, Bayes-Net
and Tree pruning classifier like J48graft and Rule learner
like Adaptive Neuro-Fuzzy Inference System has the
potential to significantly improve over the conventional
classification methods for use in medical or in general,
bioinformatics field. However there is chance of im-
provement. First, the misclassification cost is not consid-
ered explicitly in this research. In future, cost-sensitive
learning might make the study more practical and valu-
able. Second, in this survey we use only 7 rules for FIS
and ANFIS but if increasing the rules we could get more
accurate diagnosis result.
REFERENCES
[1] N. J. Nilsson, “Introduction to Machine Learning,” 2010.
http://ai.stanford.edu/~nilsson/mlbook.html
[2] M. S. Sapna and D. A. Tamilarasi, “Fuzzy Relational
Equation in Preventing Neuropathy Diabetic,” Internati-
onal Journal of Recent Trends in Engineering, Vol. 2, No.
4, 2009, p. 126.
[3] L. Carnimeo and A. Giaquinto, “An Intelligent System
for Improving Detection of Diabetic Symptoms in Retinal
Images,” IEEE International Conference on Information
Technology in Biomedicine, Ioannina, 26-28 October 2006.
[4] R. Radha and S. P. Rajagopalan, “Fuzzy Logic Approach
for Diagnosis of Diabetes,” Information Technology Jour-
nal, Vol. 6, No. 1, pp. 96-102.
doi:10.3923/itj.2007.96.102
[5] P. Jeatrakul and K. W. Wong, “Comparing the Perform-
ance of Different Neural Networks for Binary Classifica-
tion Problems,” The 8th International Symposium on Na-
tural Language Processi ng, Bangkok, 20-22 October 2009,
pp. 111-115. doi:10.1109/SNLP.2009.5340935
[6] Q. Q. Zhou, M. Purvis and N. Kasabov, “Membership
Function Selection Method for Fuzzy Neural Networks,”
University of Otago, Dunedin, 2007.
http://otago.ourarchive.ac.nz/handle/10523/1027
[7] T.-H. Lin and V.-W. Soo, “Pruning Fuzzy ARTMAP
Using the Minimum Description Length Principle in
Learning from Clinical Databases,” Proceedings of the
9th International Conference on Tools with Artificial In-
telligence, Newport Beach, 3-8 November 1997, pp.
396-403.
[8] F. Ensan, M. H. Yaghmaee and E. Bagheri, “Fact: A New
Fuzzy Adaptive Clustering Technique,” The 11th IEEE
Symposium on Computers and Communications, Sardinia,
26-29 June 2006, pp. 442-447.
doi:10.1109/ISCC.2006.73
[9] UCI Machine Learning Repository.
http://www.ics.uci.edu/mlearn/MLRepository.html
[10] S. W. Purnami, A. Embong, J. M. Zain and S. P. Rahayu,
“A New Smooth Support Vector Machine and Its Appli-
cations in Diabetes Disease Diagnosis,” Journal of Com-
puter Science, Vol. 5, No. 12, pp. 1006-1011.
[11] P. Werbos, “Beyond Regression: New Tools for Predi-
ction and Analysis in the Behavioral Sciences,” Ph.D.
Thesis, Harvard University, Cambridge, 1974.
[12] G. H. John and P. Langley, “Estimating Continuous Dis-
tributions in Bayesian Classifiers,” Proceedings of the
11th Conference on Uncertainty in Artificial Intelligence,
San Francisco, 1995, pp. 338-345.
[13] J. Quinlan, “C4.5: Programs for Machine Learning,” Mo-
rgan Kaufmann, San Mateo, 1993.
Comparison of Various Classification Techniques Using Different Data Mining Tools for Diabetes Diagnosis 97
[14] I. H. Witten and E. Frank, “Data Mining: Practical Ma-
chine Learning Tools and Techniques,” 2nd Edition,
Morgan Kaufmann, San Francisco, 2005.
[15] The Mathworks-Fuzzy Logic Toolbox, 2006.
http://www.mathworks.ch/access/helpdeskr13/help/toolb
ox/fuzzy/fuzzy.html
[16] Jang and J.-S. Roger, “Anfis: Adaptive-Network-Based
Fuzzy Inference System,” IEEE Transactions on Systems,
Man, and Cybernetics, Vol. 23, No. 3, 1993, pp. 665-685.
doi:10.1109/21.256541
[17] J. W. Han and M. Kanber, “Data Mining Concept and
Techniques,” Morgan Kaufmann Publishers, Burlington,
2000.
[18] Kappa Statistic.
http://www.dmi.columbia.edu/homepages/chuangj/kappa
Copyright © 2013 SciRes. JSEA