International Journal of Computational Intelligence and Applications

World Scientific Publishing Company

1
INSTRUCTIONS FOR TYPESETTING MANUSCRIPTS
USING MSWORD
*

FIRST AUTHOR

University Department, University Name, Address,
City, State ZIP/Zone, Country

firstauthor_id@domain_name
SECOND AUTHOR
Group, Laboratory, Address,
City, State ZIP/Zone, Country
secondauthor_id@domain_name
Received Day Month Day
Revised Day Month Day
Clustering intends to divide the subset and group the similar objects with respect to the given
similarity measure. Clustering includes number of techniques which includes statistics, pattern recognition, data
mining and other fields. Projected clustering screens the data set into numerous disjoint clusters with the
outliers so that each cluster exists in a subspace. The majority of the clustering techniques are considerably
incompetent in providing the objective function. To overcome the attribute relevancy and the redundancy by
providing the objective function, we are going to implement a new technique termed Multi cluster Dimensional
Projection on Quantum Distribution (MDPQD). This technique evolved a discrete dimensional projection
clusters using the quantum distribution model. Multi cluster dimension on quantum distribution considers the
problem of relevancy of attribute and redundancy. An analytical and empirical result offers a multi cluster
formation based on objective function and to evolve a dimensional projection clusters. Performance of the Multi
cluster Dimensional Projection on Quantum Distribution is measured in terms of an efficient multi cluster
formation with respect to the data set dimensionality, comparison of accuracy with all other algorithms and
scalability of quantum distribution.
Keywords: Dimensional projection, Discrete, Interference data, Quantum distribution model, Redundancy,
Relevancy attributes, Multi cluster.

1. INTRODUCTION

Clustering is an accepted data mining technique for a diversity of applications. One of the
incentives for its recognition is the capability to work on datasets with minimum or no a
prior knowledge. This builds clustering realistic for real world applications. In recent
times, discrete dimensional data has awakened the interest of database researchers due to
its innovative challenges brought to the community. In discrete dimensional space, the

*
For the title, try not to use more than 3 lines. Typeset the title in 10 pt Times Roman, uppercase and boldface.

Typeset names in 10 pt Times Roman, uppercase. Use the footnote to indicate the present or permanent
address of the author.

State completely without abbreviations, the affiliation and mailing address, including country. Typeset in 8 pt
Times Italic.
2 Authors Names

space from a report to its adjacent neighbor can approach its space to the outermost
reports. In the circumstance of clustering, the problem causes the space between two
reports of the same cluster to move toward the space between two reports of different
clusters. Traditional clustering methods may not succeed to distinguish the precise
clusters using quantum distribution model.

Fuzzy techniques have been used for handling vague boundaries of arbitrarily
oriented clusters. However, traditional clustering algorithms tend to break down in high
dimensional spaces due to inherent sparsity of data. Charu Puri., and Naveen Kumar.,
2011 propose a modification in the function of Gustafson-Kessel clustering algorithm for
projected clustering and prove the convergence of the resulting algorithm. It present the
results of applying the proposed projected Gustafson-Kessel clustering algorithm to
synthetic and UCI data sets, and also suggest a way of extending it to a rough set based
algorithm.

As in the case of traditional clustering, the purpose of discrete dimensional
projected clustering algorithms is to form clusters with most encouraging quality.
However, the traditional functions used in estimate the cluster quality may not be
appropriate in the predictable case. The algorithms will consequently be likely to select
few attributes values for each cluster, which might be inadequate for clustering the
reports correctly. In some previous works on projected clustering (Mohamed Bouguessa.,
and Shengrui Wang., 2009), the clusters are evaluated by using one or more of the
following criteria:

Space between the values of the attribute to produce the relevance result
Size of the selected attribute value in the cluster
Size of the associate reports in the cluster

Spontaneously, a small standard space between attribute values in a cluster
indicates that the associate reports agree on a small range of values, which can make the
reports easily restricted. A large number of selected attributes value towards the reports
are analogous at a discrete dimensional, so they are very credible to belong to the same
real cluster. Finally, a large number of reports in the cluster point out there are a high
support for the selected attributes value, and it is improbable that the small distances are
merely by chance.

All these are indicators for a high-quality multiple clusters, but there is actually
a tradeoff between them. Assume a given set of reports, it selects only attributes value
that tend to create the common space among reports, fewer attributes will be selected.
Similarly for a space obligation, locating more reports into a cluster will probably
increase the average number of attribute value chosen.

Its important to point out that in this work; we focus on Multi cluster
Dimensional Projection on Quantum Distribution (MDPQD) under the situation of
quantum distribution model to determine a discrete dimensional clustering. This form of
multi cluster is unusual from all other clustering schemes. Thus we compare our Multi
Instructions for Typing Manuscripts (Papers Title) 3


cluster Dimensional Projection on Quantum Distribution with partitioned distance-based
projected clustering algorithm and enhanced approach for projecting clustering in terms
of efficient multi cluster formation with respect to the data set dimensionality,
comparison of accuracy with all other algorithms. An analytical and empirical result
offers a multi cluster formation based on objective function and to evolve a dimensional
projection clusters.

We provide here an overview of Multi cluster Dimensional Projection on
Quantum Distribution. The rest of this paper is arranged as follows: Section 2 introduces
architecture diagram of the proposed scheme. Section 3 describes about proposed
method; Section 4 shows the evolution and experimental evaluation; Section 5 evaluated
the results and discuss about it. Section 6 describes conclusion and prospect.

2. LITERATURE REVIEW

Most existing clustering algorithms become substantially inefficient if the
required similarity measure is computed between data points in the full-dimensional
space. To address this problem, Bouguessa, M., and Shengrui Wang., 2009 a number of
projected clustering algorithms have been proposed. However, most of them encounter
difficulties when clusters hide in subspaces with very low dimensionality.

Satish Gajawada., and Durga Toshniwal., 2012 propose VINAYAKA, a semi-
supervised projected clustering method based on DE. In this method DE optimizes a
hybrid cluster validation index. Subspace Clustering Quality Estimate index (SCQE
index) is used for internal cluster validation and Gini index gain is used for external
cluster validation in the proposed hybrid cluster validation index. Proposed method is
applied on Wisconsin breast cancer dataset.

Hierarchical clustering is one of the most important tasks in data mining.
However, the existing hierarchical clustering algorithms are time-consuming, and have
low clustering quality because of ignoring the constraints. In this paper, GuoYan Hang.,
et.Al., 2009, a Hierarchical Clustering Algorithm based on K-means with Constraints
(HCAKC) is proposed.

B Shanmugapriya and M Punithavalli., 2012., an algorithm called Modified
Projected K-Means Clustering Algorithm with Effective Distance Measure is designed to
generalize K-Means algorithm with the objective of managing the high dimensional data.
The experimental results confirm that the proposed algorithm is an efficient algorithm
with better clustering accuracy and very less execution time than the Standard K-Means
and General K-Means algorithms.

Survey by Hans peter kriegel., et.Al., 2009, tries to clarify: (i) the different
problem definitions related to subspace clustering in general; (ii) the specific difficulties
encountered in this field of research; (iii) the varying assumptions, heuristics, and
intuitions forming the basis of different approaches; and (iv) how several prominent
solutions tackle different problems. Rahmat Widia Sembiring., et.Al., 2010, PROCLUS
4 Authors Names

performs better in terms of time of calculation and produced the least number of un-
clustered data while STATPC outperforms PROCLUS and P3C in the accuracy of both
cluster points and relevant attributes found.

Inspired from the recent developments on manifold learning and L1-regularized
models for subset selection, Deng Cai Chiyuan., and Zhang Xiaofei He., 2010 propose in
a new approach, called Multi-Cluster Feature Selection (MCFS), for unsupervised feature
selection. Specifically, we select those features such that the multi-cluster structure of the
data can be best preserved. The corresponding optimization problem can be efficiently
solved, since it only involves a sparse eigen-problem and a L1-regularized least squares
problem.

Yun Yang., and Ke Chen., 2011., proposed weighted clustering ensemble
algorithm provides an effective enabling technique for the joint use of different
representations, which cuts the information loss in a single representation and exploits
various information sources underlying temporal data but does not contain the extracted
feature. Jung-Yi Jiang., et.Al., 2011., have one extracted feature for each cluster. The
extracted feature, corresponding to a cluster, is a weighted combination of the words
contained in the cluster. By this algorithm, the derived membership functions match
closely with and describe properly the real distribution of the training data.

Yanming Nie., et.Al., 2012., employs a probabilistic algorithm to estimate the most likely
location and containment for each object. By performing such online inference, it enables
online compression that recognizes and removes redundant information from the output
stream of this substrate. We have implemented a prototype of our inference and
compression substrate and evaluated it using both real traces from a laboratory warehouse
setup and synthetic traces emulating enterprise supply chains.

To evolve a discrete dimensional projection clusters, a new technique named Multi
cluster Dimensional Projection on Quantum Distribution (MDPQD) scheme is presented.

3. PROPOSED MULTI CLUSTER DIMENSIONAL PROJECTION USING
QUANTUM DISTRIBUTION MODEL

The proposed work is efficiently designed for projecting the clusters in discrete
dimensional by adapting the Multi cluster Dimensional Projection on Quantum
Distribution (MDPQD). The proposed Multi cluster Dimensional Projection on Quantum
Distribution is processed under different input, intermediate and output processes.

- The input unit takes the Habitats of Human data based on Socio Democratic
Cultures (HHSD) dataset.

- The objective functions are choosing to obtain the precise goal by eliminating
the relevancy and redundancy of attributes.

Instructions for Typing Manuscripts (Papers Title) 5


- The quantum distribution technique is used to discover a model without the
relevancy analysis and redundancy of attributes.

- The activity formed in the proposed model is the multi cluster formations using
the different set if attribute values.

- The output unit contains the discrete dimensional projection clusters.

The architecture diagram of the proposed Multi cluster Dimensional Projection
on Quantum Distribution (MDPQD) is shown in Fig 3.1.






















Fig 3.1: Proposed Multi cluster Dimensional Projection on Quantum Distribution
(MDPQD) process

Habitats of
Human data
based on Socio
Democratic
Cultures (HHSD)
Dataset
Data Collection
INPUT

Objective
Function
DISCOVERY
Using Quantum Distribution
Model
TRACKING
ACTIVITIES
Multi cluster
formation with
attribute value
OUTPUT

Precise Dimensional Projection Clusters
CAUSE
6 Authors Names

In the above Figure while generating a HHSD dataset, the size of each cluster and the
domain of each attribute were first determined arbitrarily. This dataset is used to list the
different types of attributes (i.e) food habits, culture, weather conditions, business
conditions. The values are assigned to the attributes depending on the environment. The
similar attribute values are grouped together to form the clusters. Each cluster, then
precise in selecting the attribute value. The multiple clusters are formed for the different
types of attributes. For each attribute value of a cluster, a confined mean was chosen
precisely from the domain. Each report in the multi cluster determines whether to follow
the precise attribute values according to the data error rate.

Quantum distribution is optimized with a given constraints and with variables
that need to be minimized or maximized using programming techniques. An objective
function can be the result of an effort to communicate a business goal in mathematical
terms for use in assessment analysis and optimization studies. Applications are frequently
necessitate a QD model by simpler form, so that an available computational objective
function approach can be used. An a priori bound is resultant on the quantity of
error which such an rough calculation can encourage. This leads to a natural criterion
for selecting the most excellent precise attribute value is chosen.

3.1 Objective Function using QDM

We assume that the Habitats of Human data based on Socio Democratic
Cultures (HHSD) dataset consists of a set of incoming reports which are denoted by

1 Y . . .

Yi . It is assumed that the data point

Yi is received at the time stamp Si. It is

assumed that the discrete dimensionality of the HHSD data set h. The h
dimensions of the report

Yi are denoted by (y
1
i. . . y
d
i). In addition, each data point
has an error associated with the dissimilar dimensions. The error associated with the
kth dimension for data point Yi is denoted by k (

Yi ).

We remind that many indecisive data mining algorithms use the probability
density function in order to characterize the underlying behavior. Consequently, we
construct the more modest assumption that only error variances are available. Since
the diverse dimensions of the data may replicate diverse quantities, they may
correspond to very different scales. In order to take the precise behavior of the
different discrete dimensions into account, we need to perform quantum distribution
across different discrete dimensions. This is done by maintaining global statistics.
This statistics is used to compute global variances. These variances are used to scale
the data over time with values.

In order to include the greater importance of recent data points in a
developing stream, we use the concept of an objective function f(s), which quantifies
the relative importance of the different data points over time. The objective function
is drawn from the range (0, 1), and serves as a quantize factor for the relative
importance of a given data point. This function is a decreasing function, and
represents the objective of importance of a data point over time.
Instructions for Typing Manuscripts (Papers Title) 7



A commonly used objective function is the exponential objective function.
This function is defined as follows. The exponential objective function f(s) with
parameter is defined as follows as a function of f(s) = 2s (1). We note that the
value of f(s) reduces by a factor of 2 every 1/ time units. This corresponds to the
half-life of the function f(s).

3.2 Dimension Projection Clustring process

By the concept, it notify that the Discrete Dimensional Projection Clusters
do not depend on consumer limitation in determining the attribute value of each
cluster. It strives to maximize the quantum distribution technique for the each selected
attribute value and the number of selected attributes of each cluster at the same time.
As confer previously, when it indicates the superiority of a dimensional projected
cluster it maximizes the accuracy by eliminating the redundancy. It is the inclusion
process of multi clustering that allows us to implement a dynamic threshold
adjustment scheme that maximizes the criteria simultaneously.

There are two thresholds in the dimensional projections as
min
B and Q
min

that restricts the smallest amount of selected attributes for each cluster and the
minimum quantum attribute values of them. An attribute value is selected by a cluster
if and only if its quantum index with respect to the cluster is not less than Q
min.
Under
this MDPQD scheme, if an attribute value is not selected by either of two clusters, it
will be selected by the new cluster formed by merging them. However, if an attribute
value is selected by only one cluster, so that we can obtain a precise result to the
query depending on the variance of the mixed set of values at the attribute. Two
clusters are allowed to merge if and only if the resulting cluster has utmost
min
B
selected attributes.

Initially, both thresholds are set at the highest possible values, so that all
allowed merges are very feasible to engage reports from the analogous real cluster. At
some point, there will be no more talented merges with the current attribute values.
This signals the algorithm to release the values and establish a new round of
integration. The process repeats until no more merging is possible, or a target number
of clusters are reached with the attribute value. By vigorously adjusting the threshold
values in response to the merging process, the number and relevance of the selected
attributes are both exploit.

3.2.1 Algorithm for Multi cluster Dimensional Projection on Quantum
Distribution (MDPQD)

The below describes the steps to be performed

h: Habitats of Human data based on Socio Democratic Cultures (HHSD) dataset
dimensionality
min
B : Minimum number of selected attributes value per cluster
8 Authors Names

Q
min
: Minimum number of Quantum index of a selected attribute
k
th
dimensions of dataset
Begin

Step 1: Each report in a cluster
Step 2: For pace: = 0 to d-1 do
{
Step 3:
min
B : = d - pace
Step 4: Rmin: = 1 pace = (d -1)
Step 5: Foreach cluster C
Step 6: SelectAttrisVal(C, Q
min
)
Step 7: BuildQDmodel(
min
B , Q
min
)
Step 8: While Quantizeresult
{
Step 9: MC1 and MC2 are multiple clusters formed with objfunc
Step 10: Various attribute value which forms the new cluster Cn
Step 11: Cn: = MC1, MC2MCn
Step 12: SelectAttrisVal(Cn, Q
min
)
Step 13: Update Quantize result
Step 14: If clusters remained = k
Step 15: Goto 16
}
}
Step 16: Output result
End

To form the multiple clusters, each cluster keeps an objective function that
place the attribute value between the each clusters, and the best achieve is propagated.
After calculating the attribute value of all other clusters, the information of the best
cluster will be extracted from the dataset using the quantum distribution model. The
entries involving the two clusters with the same attribute value will be removed and
the value between the clusters will be inserted into the clusters. The process repeats
until no more possible merges exist, and a new clustering step will begin by achieving
the multi clustering concept.
4. EXPERIMENTAL EVALUATION
In habitats of human data based on social and demographic culture dataset,
clusters can be formed in discrete dimensions. Only a discrete dimension of attributes
is precise to each cluster, and each cluster can have a different set of precise attribute
value. An attribute is precise to a cluster if it helps identify the member reports of it.
This means the values at the precise attributes are distributed around some specific
values in the cluster, while the reports of other clusters are less likely to have such
values. Determining the multi clusters and their precise attribute value from a HHSD
dataset is known as the discrete dimensional projected clusters.
For each cluster, a discrete dimensional projected clustering algorithm
determines a set of attributes. It assumed to be more precise to the users. Discrete
Dimensional Projected clustering is potentially useful in grouping the clusters and
Instructions for Typing Manuscripts (Papers Title) 9


forms the multiple clusters using attribute value. In these datasets, the habitat levels of
different human being is taken as samples and recorded. We can view the cultural
level of different peoples as attributes of the samples or it is also taken as the samples
as the attributes of different culture people.
Clustering can be performed on attribute value of each sample. A set of precise
attribute value might co express simultaneously in only some samples. Alternatively,
a set of precise attribute value samples may have only some of the culture habits are
co expressed simultaneously. Identifying the precise attributes using the objective
function may help to improve the multi clustering objective. The selected attributes
may also suggest a smaller set of habitat for researchers to focus on, possibly reduces
the efforts spent on expensive natural experiments. In this section, we develop a
progression of experiments considered to estimate the correctness of the proposed
algorithm in terms of

i) Comparison of accuracy,
ii) Execution Time,
iii) Multi cluster formation efficiency.

5. RESULTS AND DISCUSSION
In this work we have seen how the clusters have been projected in discrete
dimensional projected spaces. The below table and graph describes the performance
of the proposed Multi cluster Dimensional Projection on Quantum Distribution
(MDPQD). In the consequence, we compared MDPQD against partitioned distance-
based projected clustering algorithm (PCKA) and Enhanced Approach for Projecting
Clustering (EAPC), in terms of accuracy.

No. of clusters Data Accuracy (%)
Proposed MDPQD EAPC method Existing PCKA
10 90 82 65
20 89 83 68
30 88 85 69
40 86 81 70
50 89 79 69
60 92 78 71
70 93 80 73
80 94 81 75
90 95 82 75
100 95 83 74
Table 5.1 No. of clusters vs. Data Accuracy
10 Authors Names


The above table (table 5.1) describes the average data accuracy based on the
number of clusters formed with respect to the HHSD dataset. The dimensionality of the
cluster of the proposed Multi cluster Dimensional Projection on Quantum Distribution
(MDPQD) is compared with an existing partitioned distance-based projected clustering
algorithm (PCKA) and Enhanced Approach for Projecting Clustering (EAPC).



Fig 5.1 No. of clusters vs. Data Accuracy
Fig 5.1 describes the average data accuracy based on the number of clusters partitioned
with respect to the dataset. The set of experiments was used here to examine the impact
of accuracy in the Multi cluster Dimensional Projection on Quantum Distribution
algorithm. MDPQD is capable to accomplish vastly precise results and its performance is
normally reliable. As we can see from Fig. 5.1, MDPQD is more scalable and accuracy in
cluster formation than the existing EAPC and PCKA algorithm. If the average clusters
accuracy is very low, only in the EAPC and PACKA by providing unsatisfactory results.
Experiments showed that the proposed MDPQD algorithm efficiently identifies the
clusters using the objective function and its dimensions precisely in a variety of
situations.

MDPQD eradicates the choice of inappropriate dimensions in all the data sets
used for experiments. This can be achieved by the fact that MDPQD initiates its process
by detecting all the regions and their positions in every dimension, facilitating it to
control the calculation of the discrete dimensions. Compared to an existing PCKA and
EAPC, the proposed MDPQD achieved better accuracy and the variance is 30-40% high.
0
10
20
30
40
50
60
70
80
90
100
10 20 30 40 50 60 70 80 90 100
No.of clusters
D
a
t
a

A
c
c
u
r
a
c
y

(
%
)
Proposed MDPQD EAPC method Existing PCKA
Instructions for Typing Manuscripts (Papers Title) 11



Discrete Data
Dimensionality
Execution Time (sec)
Proposed MDPQD EAPC method Existing PCKA
50 2 8 4
100 3 11 7
150 5 14 11
200 8 18 12
250 11 22 14
300 12 24 16
350 14 25 17
400 17 32 25
450 20 45 33
500 22 49 45

Table 5.2 Discrete Data Dimensionality vs. Execution Time
The above table (table 5.2) describes the presence of time taken to execute based
on the discrete data dimensionality partitioned with respect to the HHSD dataset. The
execution time of the cluster of the proposed Multi cluster Dimensional Projection on
Quantum Distribution (MDPQD) is compared with an existing partitioned distance-based
projected clustering algorithm (PCKA) and Enhanced Approach for Projecting Clustering
(EAPC).

Fig 5.2 Discrete Data Dimensionality vs. Execution Time
Fig 5.2 describes the presence of time to execute based on the discrete data
dimensionality with respect to the HHSD dataset. As observed from the figure 5.2,
MDPQD exhibit reliable performance from the first set of experiments on data sets with
lesser execution time taken. In tricky cases, MDPQD presents much improved results
0
10
20
30
40
50
60
50 100 150 200 250 300 350 400 450 500
Discrete Data Dimensionality
E
x
e
c
u
t
i
o
n

T
i
m
e

(
s
e
c
)
Proposed MDPQD EAPC method Existing PCKA
12 Authors Names

than the existing EAPC and PCKA method. The results stated in Fig. 5.2 recommend that
the proposed MDPQD is more interested to the proportion of data sets in execution time
parameter.

Fig 5.3 describes the consumption of time to perform the dimensional projection
of clustered based on the quantum distribution model described in the dataset. The
proposed MDPQD balances linearly with the increase in the data dimensionality. As
specified in the scalability experiments with respect to the data set size, the execution
time of MDPQD is generally provides improved results than that of EAPC and PCKA
when the time required to project the clusters in discrete dimensionality employed for
regular runs is also included.

The time consumption is measured in terms of seconds. Compared to the
existing PCKA and EAPC, the proposed MDPQD consumes less time since it gives
better cluster dimensionality result and the variance in time consumption is
approximately 30-40% low in the proposed MDPQD.

Dataset
Dimensionality
Multi Cluster formation Efficiency (%)
Proposed MDPQD EAPC method Existing PCKA
100 90 76 70
200 91 78 71
300 92 79 73
400 93 80 75
500 95 81 76
600 96 82 77
700 97 83 78
800 98 85 80
900 95 86 81
1000 96 87 82

Table 5.3 Dataset Dimensionality vs. Multi Cluster formation Efficiency

The above table (table 5.3) describes the multi cluster formation efficiency with
respect to the dataset dimensionality. The multi cluster formation efficiency for the
proposed Multi cluster Dimensional Projection on Quantum Distribution (MDPQD) is
Instructions for Typing Manuscripts (Papers Title) 13


compared with an existing partitioned distance-based projected clustering algorithm
(PCKA) and Enhanced Approach for Projecting Clustering (EAPC).



Fig 5.3 Dataset Dimensionality vs. Multi Cluster formation Efficiency

Fig 5.3 describes the multi cluster formation with the help of the objective
function. This function forms the cluster with the precise attribute values. The different
types of attributes with different set of values are used to achieve a multi cluster.
Compared to an existing EAPC and PCKA, the proposed MDPQD provides the efficient
multi cluster formation and the variance in is approximately 20-25% high in the proposed
MDPQD.

6. CONCLUSION

In this work, we efficiently achieve the multi clustering concept in Habitats of
Human data based on Socio Democratic Cultures (HHSD) dataset by professionally
introducing the proposed Multi cluster Dimensional Projection on Quantum Distribution
(MDPQD) model. The proposed scheme describes the quantum distribution model by
analyzing the data and rectify the redundancy occur on the attribute value in the dataset.
We compared MDPQD with partitioned distance-based projected clustering algorithm
and Enhanced Approach for Projecting Clustering, in terms of accuracy, and multi cluster
formation efficiency. Our experimental evaluations showed that dimensional projection
clusters considerably outperforms objective function especially in multi clustering. The
0
20
40
60
80
100
120
1
0
0
3
0
0
5
0
0
7
0
0
9
0
0
Dataset Dimensionality
M
u
l
t
i

C
l
u
s
t
e
r

f
o
r
m
a
t
i
o
n

E
f
f
i
c
i
e
n
c
y

(
%
)
Proposed MDPQD
EAPC method
Existing PCKA
14 Authors Names

experimental results showed that the proposed MDPQD scheme for the sensitive data
attributes worked efficiently by improving 25 35 % scalability and less execution time.
We show that the use of objective function in the computations can significantly improve
the quality of the underlying results.


1. Major Headings
Major headings should be typeset in boldface with the first letter of important words
capitalized.
1.1. Sub-headings
Sub-headings should be typeset in boldface italic and capitalize the first letter of the first
word only. Section number to be in boldface roman.
1.1.1. Sub-subheadings
Typeset sub-subheadings in medium face italic and capitalize the first letter of the first
word only. Section numbers to be in roman.
1.2. Numbering and spacing
Sections, sub-sections and sub-subsections are numbered in Arabic. Use double spacing
before all section headings, and single spacing after section headings. Flush left all
paragraphs that follow after section headings.
1.3. Lists of items
Lists may be laid out with each item marked by a dot:

- item one,
- item two,
- item three.

Alternatively items may also be numbered in lowercase Roman numerals:

(i) item one,
(ii) item two,
(a) lists within lists can be numbered with lowercase alphabets,
(b) second item.
(iii) item three,
(iv) item four.
Instructions for Typing Manuscripts (Papers Title) 15



or

(1) item one,
(2) item two,
(a) lists within lists can be numbered with lowercase alphabets,
(b) second item.
(3) item three,
(4) item four.
2. Equations
Displayed equations should be numbered consecutively in the paper, with the number set
flush right and enclosed in parentheses.

( ) ( )
( )
1
0
1 ,
( , )
1 ( )

=
=
< =
=
=

}
o

o o
i i
i
t
d t N d n
n t
N n d
. (1)
Equations should be referred to in abbreviated form, e.g. Eq. (1) or (2). In multiple-
line equations, the number should be given on the last line.
Displayed equations are to be centered on the page width. Standard English letters
like x are to appear as x (italicized) in the text if they are used as mathematical symbols.
Punctuation marks are used at the end of equations as if they appeared directly in the text.

Theorem 3.1. Theorems, lemmas, etc. are to be numbered consecutively in the paper.
Use double spacing before and after theorems, lemmas, etc.

Proof. Proofs should end with
3. Illustrations and Photographs
Figures are to be inserted in the text nearest their first reference. Please send one set of
originals with copies. If the author requires the publisher to reduce the figures, ensure that
the figures (including letterings and numbers) are large enough to be clearly seen after
reduction. If photographs are to be used, only black and white ones are acceptable.
Figures are to be sequentially numbered in Arabic numerals. The caption must be
placed below the figure. Typeset in 8 pt Times Roman with line spacing of 10 pt. Use
double spacing between a caption and the text that follows immediately.
16 Authors Names

Previously published material must be accompanied by written permission from the
author and publisher.
4. Tables
Tables should be inserted in the text as close to the point of reference as possible. Some
space should be left above and below the table. Tables should be numbered sequentially
in the text in Arabic numerals. Captions are to be centralized above the tables. Typeset
tables and captions in 8 pt Times roman with line spacing of 10 pt.
Table 1. This is the caption for the table. If the caption is less than one line then it is
centered. Long captions are justified to the table width manually.
Schedule Capacity Level
Business plan Financial planning
a
Planning
Production planning Resource requirement plan (RRP)
Final assembly schedule Capacity control
Master production schedule (MPS) Rough cut capacity plan (RCCP)
Stock picking schedule Inventory control
Order priorities Factory order control Execution
Scheduling Machine (work-centre) control
Operation sequencing Tool control
b

Note: Table notes.
a
Table footnote A.
b
Table footnote B.

If tables need to extend over to a second page, the continuation of the table should be
preceded by a caption, e.g., Table 1 (Continued).
5. Citation

Fig. 1. This is the caption for the figure. If the caption is less than one line then it needs to be manually
centered.
Instructions for Typing Manuscripts (Papers Title) 17


Reference citations in the text are to be numbered consecutively in Arabic numerals, in
the order of first appearance. They are to be typed in superscripts after punctuation
marks, e.g.

(1) in the statement.
1

(2) have proven
2
that this equation

When the reference forms part of the sentence, then they are typed as:

(1) One can deduce from Ref. 3 that
(2) See Refs. 13, 5 and 7 for more details.
6. Running Heads
Please provide a shortened running head (not more than eight words) for the title of your
paper.
7. Footnotes
Footnotes should be numbered sequentially in superscript lowercase Roman letters.
a

Acknowledgments
This section should come before the References. Funding information may also be
included here.
Appendix A. This is the Appendix
Appendices should be used only when sophisticated technical details are crucial to be
included in the paper.
A.1. This is the subappendix
They should come before the References. Number displayed equations occurring in the
Appendix in this way, e.g., (A.1), (A.2), etc.

2
eff
( ) 1
R H
k
S f
nV R f
o
~ . (A.1)
A.1.1. Sub-subappendix
This is the sub-subappendix.
Appendix B. Another Appendix
If there is more than one appendix, number them alphabetically.

a
Footnotes should be typeset in 8 pt Times Roman at the bottom of the page.
18 Authors Names


( ) ( )
( )
o

o o

=
=
< =
=
=

}
1
0
1 ,
( , )
1 ( )
i i
i
t
d t N d n
n t
N n d
. (B.1)
References
The references section should be labeled References and should appear at the end of
the paper. References are to be listed alphabetically with Arabic numerals. For journal
names, use the standard abbreviations. Typeset references in 9 pt Times Roman. Use the
style shown in the following examples.
Journal paper
1. R. Lorentz and D. B. Benson, Deterministic and nondeterministic flow-chart interpretations, J.
Comput. Syst. Sci. 27 (1983) 400433.
2. M. Joliat, A simple technique for partial elimination of unit productions from LR(k) parsers,
IEEE Trans. Comput. 27 (1976) 753764.
Authored book
3. M. Tinkham, Group Theory and Quantum Mechanics (McGraw-Hill, New York, 1964).
4. V. F. Kiselev and O. V. Krilov, Electron Phenomenon in Adsorption and Catalysis on
Semiconductors and Dielectrics (Nauka, Moscow, 1979) (in Russian).
Edited book
5. T. Tel, in Experimental Study and Characterization of Chaos, ed. Hao Bailin (World Scientific,
Singapore, 1990), p. 149.
6. J. K. Srivastava, S. C. Bhargava, P. K. Iyengar and B. V. Thosar, in Advances in Mssbauer
Spectroscopy: Applications to Physics, Chemistry and Biology, eds. B. V. Thosar, P. K.
Iyengar, J. K. Srivastava and S. C. Bhargava (Elsevier, Amsterdam, 1983), pp. 39, 89.
Proceedings
7. A. N. Kolmogorov, Thorie gnrale des sytmes dynamiques et mcanique classique, in Proc.
Int. Congr. Mathematicians, Amsterdam, 1954, Vol. I (North-Holland, Amsterdam, 1957), pp.
315333.
References
1. R. Lorentz and D. B. Benson, Deterministic and nondeterministic flow-chart interpretations, J.
Comput. Syst. Sci. 27 (1983) 400433.
2. M. Joliat, A simple technique for partial elimination of unit productions from LR(k) parsers,
IEEE Trans. Comput. 27 (1976) 753764.
3. M. Tinkham, Group Theory and Quantum Mechanics (McGraw-Hill, New York, 1964).
4. V. F. Kiselev and O. V. Krilov, Electron Phenomenon in Adsorption and Catalysis on
Semiconductors and Dielectrics (Nauka, Moscow, 1979) (in Russian).
5. T. Tel, in Experimental Study and Characterization of Chaos, ed. Hao Bailin (World Scientific,
Singapore, 1990), p. 149.
Instructions for Typing Manuscripts (Papers Title) 19


6. J. K. Srivastava, S. C. Bhargava, P. K. Iyengar and B. V. Thosar, in Advances in Mssbauer
Spectroscopy: Applications to Physics, Chemistry and Biology, eds. B. V. Thosar, P. K.
Iyengar, J. K. Srivastava and S. C. Bhargava (Elsevier, Amsterdam, 1983), pp. 39, 89.
7. A. N. Kolmogorov, Thorie gnrale des sytmes dynamiques et mcanique classique, in Proc.
Int. Congr. Mathematicians, Amsterdam, 1954, Vol. I (North-Holland, Amsterdam, 1957), pp.
315333.