See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318077508

Application of Principal Component Analysis (PCA) to Medical Data

Article  in  Indian Journal of Science and Technology · February 2017

DOI: 10.17485/ijst/2017/v10i20/91294

CITATIONS READS
19 6,528

6 authors, including:

Naeem Ahmed Qureshi Habibullah Magsi

Sindh Agriculture University Sindh Agriculture University
28 PUBLICATIONS   60 CITATIONS    90 PUBLICATIONS   804 CITATIONS   

SEE PROFILE SEE PROFILE

Muhammad Javed Sheikh Mubeena Pathan

Sindh Agriculture University Sindh Agriculture University
21 PUBLICATIONS   148 CITATIONS    13 PUBLICATIONS   57 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Understanding cyber harassment in Pakistani context View project

Agriculture Research Sindh View project

All content following this page was uploaded by Habibullah Magsi on 27 July 2017.

The user has requested enhancement of the downloaded file.

 ISSN (Print) : 0974-6846
Indian Journal of Science and Technology, Vol 10(20), DOI: 10.17485/ijst/2017/v10i20/91294, May 2017 ISSN (Online) : 0974-5645

Application of Principal Component Analysis (PCA)

to Medical Data
Naeem Ahmed Qureshi1, Velo Suthar1, Habibullah Magsi2, Muhammad Javed Sheikh3,
Mubeena Pathan4 and Barkatullah Qureshi4
Department of Statistics, Sindh Agriculture University, Tandojam, Pakistan; qureshistat@gmail.com,
1

vsutahar@yahoo.co.uk
2
Department of Agricultural Economics, Sindh Agriculture University, Tandojam, Pakistan; hmagsi@sau.edu.pk
3
Department of Rural Sociology, Sindh Agriculture University, Tandojam, Pakistan; ruralsociologyst@gmail.com
4
Information Technology Centre, Sindh Agriculture University, Tandojam, Pakistan; mubeena2009@gmail.com,
qureshi5939@gmail.com

Abstract
Objectives: To apply Principal Component Analysis to medical data to explore the factors thought to be very important
in increasing the risk of Ischemic Heart Diseases. Methods/Statistical Analysis: PCA was performed in R-mode using
correlation and covariance for medical data. Variables pertaining to chemical tests of blood namely cholesterol, high density
lipoprotein, triglyceride, Apo protein A-1, Apo protein B, low density lipoprotein, phospholipids, total lipid, glucose and
uric acid, are undertaken to know the relationship between them and membership of group variable. Findings: The results
indicated that among these factors cholesterol, triglyceride, Apo protein B, low density lipoprotein, phospholipids, total
lipid, and uric acid were recorded to be higher in IHD group compared to those of control group. High density lipoprotein
and Apo protein A-1 were recorded to be lower in IHD group, whereas found higher in control group. Cholesterol was
highly correlated with low density lipoprotein and moderately correlated with total lipid. Cholesterol, Apo protein B and
low density lipoprotein belonged to component 1, Apo protein A-1, phospholipids and uric acid belonged to component 2,
triglycerides and total lipid belonged to component 3 and high density lipoprotein and glucose belonged to component 4.
The first four components have explained 60.67 percent component variability. Improvements/Applications: The end
results showed that the average cholesterol level, which is considered as the main risk factor of Ischemic Heart Disease,
was found higher even in control group.

Keywords: Application, Ischemic Heart Diseases, PCA

1. Introduction related with other subset of variables which are combined

Principal component analysis (PCA) is statistical tech-
nique is applicable in the situation when the researcher is
dealing with single set of variables and wants to discover
that what are the main variable(s) which are important in
the formation of a coherent factors in such a way that there
exists no correlation between these newly formed factors.
Since PCA works in a highly correlated environment so it
chooses a set of variables that are highly dependent with
each other but, at the same time, they are quite uncor-
together to form a factor. The basic idea is that these
newly formed factors drive the underlying process due to
which the variables in the data set are supposed to corre-
late with each other. The specific goals of the component
analysis are to summarize patterns of correlations among
observed variables, to reduce a large number of observed
variables to a smaller number of factors, and to provide
an operational definition (a regression equation) for an
underlying process by using observed variables. Since the
number of factors are usually far fewer than the number

*Author for correspondence

 Application of Principal Component Analysis (PCA) to Medical Data
of observed variables, so there is a considerable parsi-
mony in using the factor analysis. Furthermore, when the
scores of factors are estimated for each subject, they are
often more reliable than scores on individual observed
variables. Anderson1, developed the asymptotic theory
for principal component analysis and proved many theo-
rems about the uses of principal component analysis.
Steps in principal component analysis include select-
ing and measuring a set of variables, preparing the
correlation matrix (to perform principal component
analysis), extracting a set of factors from the correlation
matrix, determining the number of factors, (probably)
rotating the factors to increase interpretability, and finally,
interpreting the results. Although, there are relevant sta-
tistical considerations to most of these steps, an important
test of the analysis is hidden in its interpretability.
With the passage of time numerous variants of PCA
have been proposed and applied in the literature2,3.
In4 applied nonlinear principal component analysis
(NLPCA) as a generalization of traditional principal
component analysis (PCA) that allows for the detection
and characterization of low-dimensional nonlinear struc-
ture in multivariate datasets. Likewise, nearly one and a
half decade ago, Kim and Park documented that kernel
principal component analysis (PCA) had recently been
proposed as a nonlinear extension of PCA5,6. The basic
idea was to first map the input space into a feature space
via a nonlinear map and then compute the principal com-
ponents in that feature space.
Pierpaolo and Paolo reported that Principal
Component Analysis (PCA) was a famous procedure
with the objective to produce massive amounts of quan-
titative information with the help of a small quantity of
latent variables, which are known as factors7. In their
research, they proposed the extended version of the PCA
which is quite feasible in terms of its applicability while
dealing with the interval data set. The proposed method
was named as Midpoint Radius Principal Component
Analysis (MR-PCA). Taking the interval valued data set,
the method works by using both the midpoints (which
are also known as centers of the data) and the measure of
interval width (which is also known as radii, the singular
radius) information. In order to analyzed how MR-PCA
works, the results of a simulation study and two applica-
tions on chemical data were proposed.
The book, “Principal Component Analysis:
Multidisciplinary Applications” by8 presented the appli-
cations of PCA in various fields such as taxonomy,
2 Vol 10 (20) | May 2017 | www.indjst.org
biology, pharmacy, finance, agriculture, ecology, health
and architecture etc. Similarly, PCA was applied to mul-
tispectral-multimodal optical image analysis for malaria
diagnostics9 whereas Nandi et al applied PCA in medical
image processing10.
2. Methodology
The data used for this study were originally collected
by Dr. Zaffar Ali Pirzada, Professor, Chandka Medical
College Larkana, Pakistan, for his doctoral study program
at University of Sindh Jamshoro, Pakistan. After success-
ful completion of his PhD, Dr. Pirzada provided the data
set to the Department of Statistics, Sindh Agriculture
University Tandojam, Pakistan, and allowed to apply
sophisticated test to explore the factors increasing the
risk of Ischemic Heart Diseases (IHD). In the present
study, variables pertaining to chemical tests of blood are
undertaken to know the relationship between them and
membership of group variable.
2.1 Principal Component Analysis (PCA)
Principal component analysis technique was performed
in R-mode using correlation and covariance for medical
data11. Principal component analysis is based on the fol-
lowing linear model:
Yij = βi1 X 1 j + βi 2 X 2 j + ........... + βip X pj
(1)
Where i, j = 1,2,....., p
In correlation matrix based on complete data, the
sizes of some correlation place constraints on the sizes of
others. In particular,
r13 r23 = (1 − r132 )(1 − r232 ) ≤ r12 ≤ r13 r23 + (1 − r )(1 − r )
2
13
2
23
(2)
High bi-variate correlation matrix contains factors.
Due to the reason that correlation exists between two
variables, it is useful to look at frameworks of individual
correlations where match shrewd relationships are bal-
anced for impact of every single other variable.
2.2 Bartlett’s Test
In12 defined Bartlett’s test of sphericity as a notoriously
sensitive test of the hypothesis that the correlations in a
correlation matrix are zero. But because of its sensitivity
and dependence on N, the test is likely to be significant
with samples of substantial sizes even if correlations are
Indian Journal of Science and Technology
 Naeem Ahmed Qureshi, Velo Suthar, Habibullah Magsi, Muhammad Javed Sheikh, Mubeena Pathan and Barkatullah Qureshi
very low. Therefore, use of the test is recommended only
if there are less than five cases per variable.
2.3 Anti-Image Correlation
Hostile to Image relationship framework contains the
negatives of individual relationship between’s sets of fac-
tors with impacts of different factors evacuated. On the
off chance that R is factorable, there are for the most part
little values among the off-diagonal members to diagonal
components of the anti-image matrix.
2.4 Extraction Methods
The diagonalization of the matrices can be easily traced
back in different texts related to linear algebra while hold-
ing down certain conditions. An important theorem
from matrix algebra indicates that, under certain condi-
tions, matrices can be diagonalized. The diagonalization
of a matrix leaves us with the matrix which has positive
numbers on main diagonal and all non-diagonal elements
are zeros. The elements on main diagonal represents the
variances from the correlation matrix can be written as
follows:
L = V ′RV (3)
The matrix V contains the columns which are known
as eigenvectors, while the elements on the main diagonal
of L matrix are called eigenvalues. There exists one-to-
one correspondence between the eigenvectors and the
eigenvalue i.e., each eigenvector corresponds to each
eigenvalue. However, the objective of factors analysis is
to precise how variables are correlated with each other by
keeping only those factors in the final solution which have
large corresponding eigenvalues (usually factors with
eigenvalue greater than 1 are retained and the remaining
are discarded) as each eigenvalue corresponds to different
factor.
When the transpose of the matrix of eigenvectors post-
multiplied by the identity matrix (a matrix whose main
diagonal elements are 1 and all off-diagonal elements are
0). Hence this operation of pre and post multiplication
will not affect the correlation matrix up to much extent
and can be rewritten it as given below.
V ′V = I (4)
Due to the fact that correlation matrices possess the
property of diagonalization after following a very easy
way, this property makes it feasible and possible as well to
use them efficiently using the linear algebra of eigenvec-
Vol 10 (20) | May 2017 | www.indjst.org
tors and eigenvalues in factor analysis. In the end, when
these correlation matrices are diagonalized, the informa-
tion that these matrices contain is repackaged13.
The way of calculating the eigenvalues and eigen-
vectors are not easily but in fact very laborious since it
involves solving equal number of equations say p with
equal number of unknowns say p after placing some
constraints that could not be easy to perform manually.
Upon the completion of the calculation of eigen values
and eigenvectors the only remaining job that principal
component analysis has to do is simply ‘drop out’. The fol-
lowing equations (from 5 to 8) are better explaining the
procedure.
The previously used equation can be rearranged as
follows:
R = VLV (5)
The product of three matrices jointly forms the cor-
relation matrix i.e., the product of matrices containing the
eigenvalues and the eigenvectors and its transpose. Their
reorganization ended up by taking the square root of the
eigenvalue matrix.
R = V L LV ′ (6)
Or
(
R= V L )( )
LV ′
If V L is called A and LV ′ is Aꞌ, then
R = AA′ (7)
Put it in simple, the resulting correlation matrix is
simply formed buy the product of the matrices of eigen-
vectors and the square root of the eigenvalues as shown
in the above equation (7). This equation is treated as the
basic equation in performing factor analysis which simply
consists of the product of factor loading matrix, A, and its
transpose.
It can be easily seen from the above two equations (6)
and (7) that the major task in doing Factor Analysis (FA)
or Principal Component Analysis (PCA) is finding the
eigenvalues and the eigenvectors. After their successful
computation, there exists a straightforward way to form
the (unrotated) factor loading matrix by using the follow-
ing multiplication rule:
A = V L (8)
The correlation between factors and variables is
expressed in terms of correlation matrix. Every column
Indian Journal of Science and Technology 3
 Application of Principal Component Analysis (PCA) to Medical Data

represents the correlation between the correspond- were suffering from ischemic heart disease) and control
ing variable i.e., the first column shows the correlation group. The following table shows the descriptive statistics
between the first variable, the second column shows the of ten chemical tests of blood causing ischemic heart dis-
correlation between the second variable and so on. ease patients and control group.

2.5 Communalities, Variance and 3.1 Descriptive Statistics

Covariance
The communality for a variable is the variance accounted for
by the factors. It is the squared loading multiple correlation
for the variable as predicted from the factor. Communality
is the sum of squared loading (SSL) for a variable across
factors. To find the proportion of variance explained by
a factor can be calculated by dividing the sum of square
loadings (SSL) with the number of variables, particularly
when the rotation of factors id orthogonal. Likewise, the
covariance proportion for the particular factor is obtained
by dividing the sum of square loadings (SSL) for that par-
ticular factor with the sum of communalities.
3. Results and Discussion
In this study we discuss the results of principal compo-
nent analysis for ten chemical tests of the patients (who
Table 1. Descriptive statistics of chemical tests increasing the risk of ischemic heart disease
Variable Normal values Mean
in mg/dl
It is evident from Table 1 that the majority of total patients
who were suffering from ischemic heart disease had an
average cholesterol value of 236.07 mg/dl where as that of
control group was found to be 206.69 mg/dl. High density
lipoprotein of ischemic heart disease patients was found
to be 37.69 mg/dl whereas the control group had its aver-
age value of 51.66 mg/dl. On the basis of average value
the patients who had ischemic heart disease their triglyc-
eride level was recorded to be 158.57mg/dl while that
of the control group was found to be 129.43 mg/dl. Apo
protein A-1 found in ischemic heart disease patients was
found to be 99.46 mg/dl whereas in control group this
was recorded as 130.99 mg/dl. The patients possessing
ischemic heart disease had an average value of Apo pro-
tein B 142.42 mg/dl while it was found in control group
as 105.39mg/dl. On average basis phospholipids found in
patients of ischemic heart disease was noted to be 204.86
Std. Error of Mean

Ischemic Heart Disease/Control Ischemic Heart Disease/Control

Subject in mg/dl Subject in mg/dl

Control IHD Total IHD Control Total

Cholesterol 100-200 206.6931 236.0723 231.000 2.41680 1.55897 1.43067
High Density M 35-55 51.6634 37.6921 40.1043 .99327 .35413 .40339
Lipoprotein F 45-65
Triglycerides 50-150 129.4257 158.7937 153.723 2.88905 1.51043 1.42084
Apo Protein A-1 120-150 130.9901 99.4587 104.902 1.97712 .92246 .97005

Apo Protein 0-120 105.1386 142.4174 135.981 1.52673 1.04907 1.07791

B
Low density 0-150 124.6733 171.7376 163.612 2.20787 1.46226 1.46594
Lipoprotein
Phospholipids 120-180 181.8020 204.8595 200.878 2.89000 1.85911 1.65589

Total lipid 500-1000 771.7525 922.6736 896.617 7.83452 6.96674 6.37156

Glucose R-120-180 85.8020 88.6591 88.1658 .95658 1.41844 1.18565
Uric Acid M 3.4-7.0 5.1713 6.3913 6.1807 .08796 .06195 .05673
F.M.2.4-5.7

4 Vol 10 (20) | May 2017 | www.indjst.org Indian Journal of Science and Technology
 Naeem Ahmed Qureshi, Velo Suthar, Habibullah Magsi, Muhammad Javed Sheikh, Mubeena Pathan and Barkatullah Qureshi

Table 2. Correlation matrix of various chemical tests increasing the risk of IHD
Cholest- High Trigly- Apo Apo Low density Phospho- Total Glucose Uric
erol Density cerides Protein Protein Lipoprotein lipids lipid Acid
lipopro- A-1 B
tein
Cholesterol 1.000 -.196 .230 -.246 .275 .606 .199 .421 -.086 .057

High Density -.196 1.000 -.170 .186 -.167 -.241 -.015 -.190 .150 -.029
lipoprotein
Triglycerides .230 -.170 1.000 -.236 .154 .172 -.023 .271 -.036 -.119

Apo Protein -.246 .186 -.236 1.000 -.218 -.235 -.001 -.086 .004 -.050
A-1
Apo Protein .275 -.167 .154 -.218 1.000 .377 -.028 .212 -.103 .108
B
Low density .606 -.241 .172 -.235 .377 1.000 .307 .413 -.096 .119
lipoprotein
Phospholipids .199 -.015 -.023 -.001 -.028 .307 1.000 .118 -.023 -.014

Total lipid .421 -.190 .271 -.086 .212 .413 .118 1.000 -.055 -.087

Glucose -.086 .150 -.036 .004 -.103 -.096 -.023 -.055 1.000 -.049

Uric Acid 0.57 -.029 -.119 -.050 .108 .119 -.014 -.087 -.049 1.000

Table 3. Anti-image correlation matrix of various chemical tests increasing the risk of IHD
Cholest- High Trigly- Apo Apo Low Phosp- Total Glucose Uric
erol Density cerides Protein Protein density holipids lipid Acid
lipopro- A-1 B Lipo-
tein protein
Cholesterol .7595 .0111 -.0888 .1152 -.0259 -.4360 -.0425 -.2123 .0298 -.0238

High Density .0111 .8157 .0817 -.1107 .0375 .1067 -.0405 .0763 -.1284 .0130
Lipoprotein
Triglycerides -.088 .0817 .7346 .1792 -.0493 .0137 .0644 -.1747 .0081 .1312
Apo Protein .1152 -.1117 .1792 .7454 .1118 .0793 -.0395 -.0922 .0506 .0296
A-1
Apo Protein -0.259 .0375 -.0493 .1118 .7692 -.2550 .1456 -.0531 .0645 -.0732
B
Low density -.4360 .1067 .0137 .0793 -.2550 .7002 -.2786 -.1978 .0136 -.1145
Lip protein
Phospholipids -.0425 -.0405 .0644 -.0395 .1456 -.2786 .5752 -.0009 .0106 .0505
Total lipid -.2123 .0763 -.1747 -.0922 -.0531 -.1979 -.0009 .7828 -.0046 .1279

Glucose .0298 -.1284 .0081 .0506 .0645 .0136 .0106 -.0046 .6881 .0356

Uric Acid -.0238 .0130 .1312 .0296 -.0732 -.1145 .0505 .1279 .0356 .4974

Vol 10 (20) | May 2017 | www.indjst.org Indian Journal of Science and Technology 5
 Application of Principal Component Analysis (PCA) to Medical Data
mg/dl while those who belonged to control group their
average value was recorded to be 181.80mg/dl. Total lipid
found in ischemic heart disease patients was found to be
922.67 mg/dl whereas it was recorded in control group as
771.75 mg/dl. Average value of glucose found in ischemic
heart disease patients was noted as 88.66 mg/dl while
in control group this was found to be 85.80 mg/dl. The
average value of uric acid found in ischemic heart dis-
ease patients was found to be 6.39 mg/dl while in control
group it was recorded to be 5.17 mg/dl.
When Table 1 studied, it was found that the average
values of cholesterol, triglyceride, Apo protein B, low den-
sity lipoprotein , phospholipids, total lipid, glucose and
uric acid were higher in ischemic heart disease patients
as compared to those of control group while high density
lipoprotein and Apo protein A-1 were found to be higher
in control group.
Though the control group has the normal values but
due to increase in average value of cholesterol the people are
under heavy risk of increasing the ischemic heart disease.
3.2 Correlation Analysis
Correlation coefficients between ten chemical tests for
Ischemic Heart Disease are shown in Table 2. It could be
observed from this table that cholesterol has positively
high correlation with low density lipoprotein (0.606) and
it is moderately correlated with total lipid (0.421). There
is negative correlation between high density lipoprotein
and low density lipoprotein. The observed correlation
of triglyceride with total lipid (0.271) is positively weak.
Apo protein A-1 has negatively weak whereas Apo pro-
tein B has positively weak correlation with low density
lipoprotein. There is highly positive correlation observed
between low density lipoprotein and total lipid.
3.3 Anti Image Correlation Analysis
Discussing Table 3, it becomes clear that almost all the
values at the lower or upper half of the main diagonal are
smaller, i.e. close to zero, which indicates that variables
are relatively free from unexplained correlations. In the
main diagonal of anti-image correlation matrix, each
value describes the measure of sampling adequacy for
the respective item. Values less than 0.5 may indicate that
a variable does not seem to fit with the structure of the
variables. Table 3 all the variables in main diagonal have
values more than 0.5, which reveals that, the variables
have adequate samples.
6 Vol 10 (20) | May 2017 | www.indjst.org
3.4 Initial and Extraction Communalities
Analysis
In principal component analysis initial and extraction
communalities for variables are the variances accounted
for the factors. For variance and co-variance analysis ini-
tial communalities remain always equal to 1.0. For the
extraction communalities, these values are the proportion
to variance of each variable by the rest of the variables.
The extraction communalities are the sum of squared
loadings for a variable across factors, in Table 4, the
extraction communalities for cholesterol are (0.770)2
+ (0.180)2 + (0.0823)2 + (0.0877)2 = 0.639. That was
Factor1+Factor2+Factor3+Factor4 accounted for 63.9%
of the variance calculated for cholesterol. Variances cal-
culated for high density lipoprotein, triglyceride, Apo
protein A-1, Apo protein B, low density lipoprotein, phos-
pholipids, total lipid, glucose and uric acid were 42.1%,
57.6%, 435.4%, 74.7%, 45.5%, 64.1%, 53.6%, 81.6% and
72.3% respectively.
As already stated in the last section that those factors
which have small variances do not fit in the factor solution
and should be discarded from the analysis, our empirical
results presented in Table 4 shows that high density lipo-
protein and Apo protein B can be easily discarded from
the analysis due to explaining a very small proportion of
the variance. Since this proportion of variance explained
by a factor is calculated by dividing the sum of square
loadings with the number of variables used, hence for
these two factors, it was calculated as 99.1% whereas the
covariance for these factors was calculated as 100%.
3.5 Component Analysis
Table 5 shows the eigenvalues and total variance explained
for our factor solution. There are ten variables used in the
sample, because the aim of factor analysis is to precise
the pattern of correlation with a factor as possible. Since
every eigenvalue corresponds to a different factor (usually
only factor with large eigenvalues are retained). In a good
factor analysis, these few factors defined the whole corre-
lation matrix. When no limit is placed on the number of
factors eigenvalues are computed for each of ten possible
components as shown in Table 5.
It is quite clear from the Table 6 that the first four com-
ponents has their eigenvalues over 1 and are large enough
to be retained, their variances are 26.65%, 12.02%, 11.79%
and 10.19% respectively. These four components describe
the 60.67% of the total variance.
Indian Journal of Science and Technology
 Naeem Ahmed Qureshi, Velo Suthar, Habibullah Magsi, Muhammad Javed Sheikh, Mubeena Pathan and Barkatullah Qureshi

Table 4. Relationship among loadings, communalities, variance, co-variance of orthogonally rotated factors
Variables Initial Factor 1 Factor 2 Factor 3 Factor 4 Extraction
Cholesterol 1.000 .770 .180 0.0823 0.0877 0.639765

High density lipoprotein 1.000 -.452 .262 0.187 0.337 0.421486

Triglycerides 1.000 .447 -.545 0.282 0.0154 0.576595

Apo Protein A-1 1.000 -.448 .363 0.197 -0.376 0.512658

Apo Protein B 1.000 .548 -.133 -0.350 0.122 0.455377
Low density lipoprotein 1.000 .804 .307 -0.0146 0.0798 0.747246
Phospholipids 1.000 .276 .681 0.315 -0.0459 0.641269
Total Lipid 1.000 .633 -0.002477 0.354 -0.104 0.536827
GLUCOSE 1.000 -.198 -0.009493 0.328 0.818 0.816002
Uric Acid 1.000 0.07920 .311 -0.752 0.235 0.723723
Sum of square of loadings 2.666779 1.202354 1.180417 1.021397 6.070948
Proportion of Variance 0.266678 0.120235 0.118042 0.10214 0.607095
Proportion of co-variance 0.439269 0.19805 0.194437 0.168243 1.000000

Table 5. Component variability percentage correlation analysis by using eigenvalues

Component Initial Eigenvalues Extraction Sums of Squared Loadings
Total %of variance Cumulative % Total % of variance Cumulative %
1 2.666 26.658 26.658 2.666 26.658 26.658
2 1.202 12.022 38.680 1.202 12.022 38.680
3 1.179 11.794 50.474 1.179 11.794 50.474
4 1.020 10.196 60.670 1.020 10.196 60.670
5 .869 8.694 69.364
6 .791 7.906 77.270
7 .738 7.382 84.652
8 .649 6.494 91.146
9 .531 5.309 96.455
10 .355 3.545 100.000

Table 6. Analysis of component score coefficient matrix

Variables Components
1 2 3 4
Cholesterol .289 .149 .070 .086
High Density Lipoprotein -.169 .218 .158 .330
Triglyceride .168 -.453 .239 .015
Apo Protein A-1 -.168 .302 .167 -.369
Apo protein B .205 -.111 -.296 .120
Low Density Lipoprotein .302 .255 -.012 .078
Phospholipids .103 .566 .267 -.045
Total Lipid .237 -.002 .300 -.102
Glucose -.074 -.008 .278 .803
Uric Acid .030 .259 -.638 .226

Vol 10 (20) | May 2017 | www.indjst.org Indian Journal of Science and Technology 7
 Application of Principal Component Analysis (PCA) to Medical Data
3.6 Component Score Coefficient Analysis
The component loading matrix is a matrix of correlation
between components and variables. The first column is
the correlation between the first component and each
variables, the second column is the correlation between
the second component and each variable and so on. Table
6 reveals that there were few strong correlation observed
between components and their respective variables. A
component is interpreted from variables that are highly
correlated with it-that had loadings on it.
In the evident from Table 6 that the variables that
had high loading are assumed to belong with the com-
ponent 1, i.e., cholesterol (0.289), Apo protein B (0.205)
and low density lipoprotein (0.302). Likewise in compo-
nent 2 the variables had the high weights were found to
be Apo protein A-1 (0.302), phospholipids (0.566) and
uric acid (0.259). Similarly, triglyceride (0.239) and total
lipid (0.300) belonged to component 3. In the same way
in component 4 the variables with high scores were high
density lipoprotein (0.330) and glucose (0.803).
4. Conclusion
The step wise principal Component Analysis was used to
explore the effects of these chemical tests which are sig-
nificantly affecting the patients namely cholesterol, high
density lipoprotein, triglyceride, Apo protein A-1, Apo
protein B, low density lipoprotein, phospholipids, total
lipid, glucose and uric acid of patients who were suffer-
ing from Ischemic Heart Disease. In the case of finding
correlations between variables, it was observed that
Cholesterol was highly correlated with low density lipo-
protein and moderately correlated with total lipid. Anti
image correlations interpreted that the variables were rel-
atively free from unexplained correlation. It was observed
from communalities extractions that high density lipo-
protein and Apo protein B has small variances and do not
fit in the factor solution. To know the variability between
the components, it was found that first four components
account for exactly 60.67% of the total variability. These
four components adequately describe the whole correla-
tion matrix. Finally it was obvious to see that cholesterol,
Apo protein B and low density lipoprotein had high value
and belongs to the components 1. In component 2, it was
found that Apo protein A-1, phospholipids and uric acid
had the high scores. The variables that had high loadings
8 Vol 10 (20) | May 2017 | www.indjst.org
in the component 3 were triglyceride and total lipid. The
variables had high weight and belonged to component 4
were high density lipoprotein and glucose. The important
finding of this study is that the average cholesterol level
which is considered to be main factor increasing the risk
of Ischemic Heart Disease is found to be higher even in
control group than normal values.
5. References
1. Anderson TW. Asymptotic theory for principal component
analysis. Annals of Mathematical Statistics. 1963; 34:122-
48. Crossref.
2. Sharma UM. Hybrid Feature Based Face Verification and
Recognition System Using Principal Component Analysis
and Artificial Neural Network. Indian Journal of Science
and Technology. 2015 Jan; 8(S1):1-6. Crossref.
3. Sindhubargavi R, Sindhu MY, Saravanan R. Spectrum
Sensing Using Energy Detection Technique for Cognitive
Radio Networks Using PCA Technique. Indian Journal of
Science and Technology. 2014 Apr; 7(S4):40-5.
4. Linting M, Meulman JJ, Groenen PJ, van der Koojj AJ. Non-
linear principal components analysis: introduction and
application. Psychological Methods. 2007; 12(3):336-58.
Crossref. PMid:17784798.
5. Kim KI, Park SH. Kernel principal component analysis for
texture classification. IEEE Signal Processing Letters. 2001;
8(2):39-41. Crossref.
6. Jayabalan L, Ganapathi P. Anomaly based Malicious Traffic
Identification using Kernel Extreme Machine Learning
(KELM) Classifier and Kernel Principal Component
Analysis (KPCA). Indian Journal of Science and Technology.
2016 Apr; 9(13):1-11. Crossref.
7. D’Urso P, Paolo G. A least squares approach to Principal
Component Analysis for interval valued data. Economics
and Statistics Discussion Paper. 2002; p. 1-29.
8. Sanguansat P. Croatia: InTech Publishers: Principal
Component Analysis: Multidisciplinary Applications.
2012. Crossref.
9. Omucheni DL, Kaduki KA, Bulimo WD, Angeyo HK.
Application of principal component analysis to multi-
spectral-multimodal optical image analysis for malaria
diagnostics. Malaria Journal. 2014; 13:1-11. Crossref
.PMid:25495235 PMCid:PMC4364660.
10. Nandi D, Ashour AS, Samanta S, Chakraborty S,
Mohammad AMS, Dey N. Principal component analysis in
medical image processing: a study. International Journal of
Image Mining. 2015; 1(1):65-86.
11. Murthy HSN, Meenakshi M. Efficient Algorithm for
Early Detection of Myocardial Ischemia using PCA based
Indian Journal of Science and Technology
 Naeem Ahmed Qureshi, Velo Suthar, Habibullah Magsi, Muhammad Javed Sheikh, Mubeena Pathan and Barkatullah Qureshi

Features. Indian Journal of Science and Technology. 2016 13. Murali M. Principal Component Analysis based Feature
Oct; 9(39):1-13. Vector Extraction. Indian Journal of Science and
12. Snedecor GW, Cochran WG. USA: Iowa State University Technology. 2015 Dec; 8(35):1-4. Crossref.
Press: Statistical Methods. 8th ed. 1989.

Vol 10 (20) | May 2017 | www.indjst.org Indian Journal of Science and Technology 9

View publication stats