A Competent and Empirical Model of Distributed Clustering

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 02 | Feb-2016 www.irjet.net p-ISSN: 2395-0072
© 2015, IRJET ISO 9001:2008 Certified Journal Page 1
A Competent And Empirical Model Of Distributed Clustering
A.Vasudeva Rao
Associate Professor
Dept of CSE, Dadi Institute of Engineering Technology, Anakapalli, Visakshpatnam,JNTUK
Abstract: - Categorizing the different types of data over
network is still an important research issue in the field
of distributed clustering. There are different types of
data such as news, social networks, and education etc.
All this text data available in different resources. In
searching process the server have to gather information
about the keyword from different resources. Due to
more scalability this process leads more burdens to
resources. So we introduced a framework that consists
of efficient grouping method and efficiently clusters the
text in the form of documents. It guarantees that more
text documents are to be clustered faster.
1.INTRODUCTION
Information retrieved from a computer which is
already stored in that computer. Information is stored in the
form of documents. Computer may not be store the
information as same in the documents. It may overwrite the
data to computer understandable language. The document
may contain abstract at the starting of the document or may
the words list. It is must and should process to maintain the
document in the computer.
In the practical approaches researchers considered
that input document must contain the abstract or tile and
some text. This takes more time to process and the
document is notated with the main class of the document.
The documents stored in the computer must maintain index
with understandable classes. It is only possible that the
document is fully filtered. Such as the removing of the
grammar words and the duplicate words and maintain more
frequency terms are in the top position.
In the document indexing is the main part to
maintain the numbers of tokens such as keywords in a
language. This is only possible of clustering means that
grouping of the keywords which have same properties.
There are so many types of grouping and some of them is
serial search. Serial search is defined as the match queries
with every document in the computer and group the files to
match the query with respect to keyword and that is so
called as cluster representative.
Cluster representative will process the input query
and perform search for the documents which is matched.
Apart from that the documents which are not matched is
eliminated from the group.
There are more number of clusters algorithms are
there to cluster the documents . The ultimate goal is clustering
algorithms’ group a set of documetns into subsets clusters.The
algorithms’ goal is to
grpup similar documents and remaining documents are
deviate from the clusters.
Classification of a document into a classification
slot and to all intents and purposes identifies the document
with that slot. Other documents in the slot are treated as
identical until they are examined individually. It would
appear that documents are grouped because they are in
some sense related to each other; but more basically and
they are grouped because they are likely to be wanted
together and the logical relationship is the means of
measuring this likelihood. In this people have achieved the
logical organization in two different ways. Initially through
direct classification of the documents and next via the
intermediate calculation of a measure of closeness between
documents. The basic approach has proved theoretically to
be intractable so that any experimental test results cannot
be considered to be reliable. The next approach to
classification is fairly well documented now and there are
some forceful arguments recommending it in a particular
form. It is this approach which is to be emphasized here.
This process is used for the document matching. It
searches for the document in the clusters which is matching
to another document and the matching frequency of the
documents. Group with high score frequency which is
matching is the new document is assigned to that group. It
leads to the retrieval process slow .
Document clustering (or Text clustering) is
documents and keyword extraction and fast information
retrieval . Document clustering is the use of descriptors.
They are sets of words such as word bag that explains the
contents in the cluster. Clustering of documents considered
to be a centralized process which includes web document
clustering for search users. It is divided into two types such
as online and offline. Online clustering have efficiency
problems than offline clustering.
Most of the classifications are based on binary
relationships. These relationships of classification method
construct the system of clusters. It is explained in different
types such as similarity, association and dissimilarity.
Abort the dissimilarity it will be defined mathematically
later and the other two parameters are means the
association will be reserved for the similarity between
objects. Similarity measure is designed to find the equity
between the keywords and the documents. Possible similar
tokens are grouped together.

There are two types of algorithms such as
hierarchical based algorithm which calculations are
depends upon the links and the averages of the similarity.
Aggregation clustering is more compatible to browsing.
These two have their limitations in the efficiency. There is
another algorithm that is developed using the K-means
algorithm and its features. It has more effiency and reduces
the computations in the clustering which also gives
accurate results.
In the process of searching the user gives a
keyword to search and its displays relevant documents. The
internal process is find the similarity between or finding
the documents from the resources. For finding the
similarity we have different types of similarity measures.
Text Clustering methods are divided into three
types. They are partitioning clustering, Hierarchal
clustering, fuzzy is clustering. In partitioning algorithm,
randomly select k objects and define them as k clusters.
Then calculate cluster centroids and make clusters as per
the centroids. It calculates the similarities between the text
and the centroids. It repeats this process until some criteria
specified by the user.
Hierarchical algorithms build a cluster hierarchy;
clusters are composed of clusters that are composed of
clusters. There is a way from single documents up to the
whole text set or any part of this complete structure. There
are two natural ways of constructing such a hierarchy:
bottom-up and top-down. It puts all documents into one
cluster until some criteria reached.
In this paper we introduced new process of
clustering. In related work section briefly explained about
the traditional clustering algorithms. If text data present in
single resource, we can cluster easily because we can
gather information from centralized system. Our situation
is to cluster text data in different resources such as
decentralized systems. In this we use normal clustering
algorithms we cannot perfectly cluster the text data. So we
used clustering algorithm using some properties of the
traditional clustering algorithm and it has the capability to
use in distributed systems also. Hierarchical techniques
produce a nested sequence of partitions, with a single and
all inclusive cluster at the top and singleton clusters of
individual points at the bottom. In every intermediate level
can be viewed as combining two clusters from the next
lower level (or splitting a cluster from the next higher
level).
II. RELATED WORK
Hierarchical techniques and partitional clustering
techniques are single level division of the data points. If K
is the number of clusters given by the user the clustering
algorithm finds all K clusters. The traditional hierarchical
clustering which divide a cluster to get two clusters or
merge two clusters. Hierarchical method used to generate a
division of K clusters and the repeat the steps of a
partitional scheme can provide a hierarchical clustering.
There are a number of partitional algorithms and
only describe the K-means algorithm which is mainly used
in clustering. K-means algorithm based on centroid which
represent a cluster. K-means use the centroids and which is
the mean or median point of a group of points. Centroid is
not an actual data point. Centroid is the most important in a
cluster. The values for the centroid are the mean of the
numerical attributes and the mode of the categorical
attributes.
K-means Clustering :
Partitioned clustering method is related with a centroid and
every point is input to the cluster with the distance less to
the centroid. Cluster number can be specified by the user
only.
The basic algorithm is very simple
The basic K-means clustering technique is shown
below. We can explain it later in the following sections.
Traditional K-means Algorithm for finding K clusters.
1. Select K points as the initial centroids.
2. Assign all points to the closest centroids.
3. Re-compute the centroids of each cluster.
4. Repeat steps 2 and 3 until the centroids don’t change.
Initial centroids are often chosen randomly. The centroid is
(typically) the mean of the points in the cluster. Similarity
is measured by Euclidean distance or cosine similarity or
correlation and K-means will converge for common
similarity measures mentioned above. In the first few
iterations
Complexity is O (n * K * I * d)
n = number of points, K = number of clusters,
I = number of iterations, d = number of attributes
Similarity calculation is the main part in our
proposed work. We use cosine similarity; it is explained in
our proposed work. It means algorithm the keywords or
tokens are to be clustered up to some criteria to be reached.
A key limitation of k-means is its cluster model. It based
on clusters that are separated in a way that the mean value
related towards the center of cluster. The clusters are
expected as similar size and the assignment to the nearest
cluster center is the correct assignment.
In k-means algorithm more number of keywords
present in the document it takes more time to process and
also the computational complexity also high.
A) Initial features of Clustering
There are two types of searching such as central
servers and ﬂooding-based searching targeted scalability
and eﬃciency of distributed systems. The central servers
disqualified with a linear complexity for storage because
they concentrate all resources of data and nodes in one
single system. Flooding-based method avoids the
management of references on other nodes and they face
scalability problems in the communication process.

Distributed hash tables are the main tool for
maintaining the structure the distributed systems. It
maintains the position of the nodes in the communication
system and it has the below properties. They are
It maintains the references to the nodes and it has the
complexity O(logN) where N depicts the number of nodes
in the channel. For finding the path of nodes and data items
into address and routing to a node leads to the data items
for which a certain node is responsible. The queries given
by the user reaches the resource by small nodes in the
network to the target node. Distributing the identifiers of
nodes and equally outputs the system and reduce load for
retrieving items should be balanced among all nodes.
Not an every node maintains the individual
functionality and equally distributed the work of every
node. So distributed hash tables are considered to be very
robust against random failures and attacks.
A distributed index provides a definitive answer about
results. If a data item is stored in the system and the DHT
guarantees that the data is found.
The main initial thing is tokenizing the key words
in documents. Tokenizing means dividing the keywords,
for this we construct DHT (Distributed hash table). It
contains keyword and respective location of the keyword in
the document. It also contains frequency of the keyword in
the documents. In our work we construct DHT for clusters.
DHT provides lookup for the distributed networks
by constructing hash tables. By using DHT distributed
networks or systems maintains mapping among the nodes
in the network. It maintains more number of nodes. It is
very useful in constructing large networks.
In the traditional clustering the data points are
clustered up to some criteria reached. But our proposed
work clustering applied on all data points no point remains.
All data points should be placed in clusters.
III.PROPOSED WORK
For a given number of documents construct
distributed hash table. For every document we construct
DHT which contains tokens or terms and keys. These
tables are referenced for next clustering process.
Second is similarity between the nodes in the
network so we use cosine similarity. In this similarity
calculation we consider only the similar properties between
the edges. The reason of taking cosine similarity measure is
explained below.
Cosine similarity is a measure of similarity
between two vectors of an inner product space that
measures the cosine of the angle between them. The cosine
of 0° is 1 it is less than 1 for any other angle. It takes
magnitude of two vectors with the same orientation have a
Cosine similarity of 1, two vectors at 90° have a similarity
of 0 and two vectors opposed to have a similarity of -1 and
it is independent of their sign. This similarity is particularly
Note that these are only apply for any number of
dimensions and their Cosine similarity is most commonly
used in high-dimensional positive spaces. In Information
Retrieval and text mining and each term is notionally
assigned a different dimension and a document is
characterized by a vector where the value of each
dimension corresponds to the number of times that term
appears. Cosine similarity then gives a useful measure of
how similar two documents are likely to be in terms of
their subject matter. The technique is also used to measure
cohesion within clusters in the field of data mining.
Cosine distance is a term often used for the
complement in positive space, that is: Dc(A,B)=1-Sc(A,B) .
It is important to note and that this is not a proper distance
metric as it does not have the triangle inequality property.
The same ordering and necessary to convert to
trigonometric distance (see below.) One of the reasons for
the popularity of Cosine similarity is that it is very efficient
to evaluate especially for sparse vectors and only the non-
zero dimensions need to be considered.
In our work the cosine similarity between documents
and cluster centroids and it is defined as
Cos(d,c)=∑tεdTF(t,d)*
Next Clustering, Consider two nodes have some
documents. On these documents we perform
Aggloromative Hierarchal Clustering algorithm.
In this it follows the following steps.
Take all keys words such as data points in the document.
Cluster the points using the similarity measure, All
points placed in clusters.
Then take least distanced cluster and start index from
zero. Then merge all points in the clusters and perform
clustering process.
Order top ten clusters
For every cluster it maintains gist, keywords and the
frequency of the keywords of every cluster.
The cluster which is present in the node that referred as
cluster holder.
If new document appears, it calculates similarity
measure with every cluster. The document will place on the
highest similarity cluster.
Note that the similarity is compared for new
document is with cluster centroids and the new document.
The above generated cluster summary is used for
calculation of the similarity measure.
The experimental results shown below:
used in positive space, where the outcome is neatly
bounded in [0,1].

For new document , the calculations and the assigning is
shown above.
IV.CONCLUSION
In our proposed work we designed a method for
clustering of text in distributed systems. For increasing the
complexity of calculations our work very useful. In real
time applications also it is very helpful. For reducing the
resources work and the processing it works efficiently.
Compared to traditional process in distribution systems text
clustering process faster.
REFERENCES
[1] Y. Ioannidis, D. Maier, S. Abiteboul, P. Buneman, S.
Davidson, E. Fox, A. Halevy, C. Knoblock, F. Rabitti, H.
Schek, and G. Weikum, “Digital library information-
technology infrastructures,” Int J Digit Libr, vol. 5, no. 4,
pp. 266 – 274, 2005.
In this every node it maintains cluster summary.
The above cluster holder also maintained in the node.
[2] P. Cudr´e-Mauroux, S. Agarwal, and K. Aberer,
“Gridvine: An infrastructure for peer information
management,” IEEE Internet Computing, vol. 11, no. 5,
2007.
[3] J. Lu and J. Callan, “Content-based retrieval in hybrid
peer-topeer networks,” in CIKM, 2003.
[4] J. Xu and W. B. Croft, “Cluster-based language models
for distributed retrieval,” in SIGIR, 1999.
[5] O. Papapetrou, W. Siberski, and W. Nejdl, “PCIR:
Combining DHTs and peer clusters for efficient full-text
P2P indexing,” Computer Networks, vol. 54, no. 12, pp.
2019–2040, 2010.
[6] S. Datta, C. R. Giannella, and H. Kargupta,
“Approximate distributed K-Means clustering over a peer-
to-peer network,” IEEE TKDE, vol. 21, no. 10, pp. 1372–
1388, 2009.
[7] M. Eisenhardt, W. M¨ uller, and A. Henrich,
“Classifying documents by distributed P2P clustering.” in
INFORMATIK, 2003.
[8] K. M. Hammouda and M. S. Kamel, “Hierarchically
distributed peer-to-peer document clustering and cluster
summarization,” IEEE Trans. Knowl. Data Eng., vol. 21,
no. 5, pp. 681–698, 2009.
[9] H.-C. Hsiao and C.-T. King, “Similarity discovery in
structured P2P overlays,” in ICPP, 2003.
[10] I. Stoica, R. Morris, D. Karger, F. Kaashoek, and H.
Balakrishnan, “Chord: A scalable peer-to-peer lookup
service for internet applications,” in SIGCOMM, 2001.
[11] K. Aberer, P. Cudr´e-Mauroux, A. Datta, Z.
Despotovic, M. Hauswirth, M. Punceva, and R. Schmidt,
“P-Grid: a selforganizing structured P2P system,”
SIGMOD Record, vol. 32, no. 3, pp. 29–33, 2003.
[12] A. I. T. Rowstron and P. Druschel, “Pastry: Scalable,
decentralized object location, and routing for large-scale
peer-to-peer systems,” in IFIP/ACM Middleware,
Germany, 2001.
[13] C. D. Manning, P. Raghavan, and H. Schtze,
Introduction to Information Retrieval. Cambridge
University Press, 2008.

[14] M. Steinbach, G. Karypis, and V. Kumar, “A
comparison of document clustering techniques,” in KDD
Workshop on Text Mining, 2000.
[15] G. Forman and B. Zhang, “Distributed data clustering
can be efficient and exact,” SIGKDD Explor. Newsl., vol.
2, no. 2, pp. 34– 38, 2000.
[16] S. Datta, K. Bhaduri, C. Giannella, R. Wolff, and H.
Kargupta, “Distributed data mining in peer-to-peer
networks,” IEEE Internet Computing, vol. 10, no. 4, pp.
18–26, 2006.
[17] S. Datta, C. Giannella, and H. Kargupta, “K-Means
clustering over a large, dynamic network,” in SDM, 2006.
[18] G. Koloniari and E. Pitoura, “A recall-based cluster
formation game in P2P systems,” PVLDB, vol. 2, no. 1, pp.
455–466, 2009.
[19] K. M. Hammouda and M. S. Kamel, “Distributed
collaborative web document clustering using cluster
keyphrase summaries,” Information Fusion, vol. 9, no. 4,
pp. 465–480, 2008.
[20] M. Bender, S. Michel, P. Triantafillou, G. Weikum,
and C. Zimmer, “Minerva: Collaborative p2p search,” in
VLDB, 2005, pp. 1263– 1266.

A Competent and Empirical Model of Distributed Clustering

Recommended

Recommended

More Related Content

What's hot

What's hot (18)

Similar to A Competent and Empirical Model of Distributed Clustering

Similar to A Competent and Empirical Model of Distributed Clustering (20)

More from IRJET Journal

More from IRJET Journal (20)

Recently uploaded

Recently uploaded (20)

A Competent and Empirical Model of Distributed Clustering