Performance Analysis of Different Clustering Algorithm

IOSR Journal of Computer Engineering (IOSR-JCE)
e-ISSN: 2278-0661,p-ISSN: 2278-8727, Volume 17, Issue 1, Ver. III (Jan – Feb. 2015), PP 25-29
www.iosrjournals.org
DOI: 10.9790/0661-17132529 www.iosrjournals.org 25 | Page
Performance Analysis of Different Clustering Algorithm
1
Naresh Mathur, 2
Manish Tiwari, 3
Sarika Khandelwal
1
M.Tech Scholar, 2
Assistant Professor, 3
Associate Professor
1,2&3
Department of Computer Science Engineering, Geetanjali Institute of Technical Studies, Udaipur.
Abstract: Clustering is the process of grouping objects into clusters such that the objects from the same
clusters are similar and objects from different clusters are dissimilar. The relationship is often expressed as
similarity or dissimilarity measurement and is calculated through distance function. some of the outlier
detection techniques are distance based outlier detection distribution based outlier detection density based
outlier detection and depth based outlier detection The goal of this paper is the detection of outliers with high
accuracy and time efficiency. The methodology discussed here is able to save a large amount of time by
selecting a small subset of suspicious transactions for manual inspection which includes most of the erroneous
transactions.
Keywords: PAM.CLARA, CLARANS, ECLARANS
I. Introduction
Data mining is the method of extracting patterns from data. It can be used to uncover patterns in data
but is often carried out only on sample of data. analysis is a tool for exploring the structure of data. Cluster
analysis is the organization of a collection of patterns (usually represented as a vector of measurements, or a
point in a multidimensional space) into clusters based on similarity. Intuitively, patterns within a valid cluster
Clustering is the process of grouping objects into clusters such that the objects from the same clusters are similar
and objects from different clusters are dissimilar. The relationship is often expressed as similarity or
dissimilarity measurement and is calculated through distance function. Clustering is useful technique for the
discovery of data distribution and patterns in the underlying data. are more similar to each other than they are to
a pattern belonging to a different cluster.
Fig 1: Outlier Detection Module
Outliers detection is an outstanding data mining task, referred to as outlier mining. Outliers are objects that do
not comply with the general behavior of the data. By definition, outliers are rare occurrences and hence
represent a small portion of the data.
II. Categorization Of Clustering Techniques
According to Data Mining concepts and Techniques by Jiawai Han and Micheline Kamber clustering
algorithm partition the dataset into optimal number of clusters.
They introduce a new cluster validation criterion based on the geometric property of data partition of
the dataset in order to find the proper number of clusters. The algorithm works in two stages. The first stage of
the algorithm creates optimal number of clusters , where as the second stage of the algorithm detect outliers.
2.1 Cluster Algorithms:
Algorithms which are being used for outlier detection are-
 PAM(Partitioning around Medoids)
 CLARA(Clustering large applications)

Performance Analysis Of Different Clustering Algorithm
 CLARANS(Clustering large applications by randomized search)
 ECLARANS(Enhanced Clarans)
2.1.1 PAM (Partitioning Around Medoids) –
PAM (Partitioning Around Medoids) was developed by Kaufman and Rousseeuw. To find k clusters,
PAM’s approach is to determine a representative object for each cluster. This representative object, called a
medoid, is mean to be the most centrally located object within the cluster. Once the Medoids have been selected,
each non selected object is grouped with the medoid to which it is the most similar.
Procedure-
1. Input the dataset D
2. Randomly select k objects from the dataset D
3. Calculate the Total cost T for each pair of selected Si and non selected object Sh
4. For each pair if T si < 0, then it is replaced Sh
5. Then find similar medoid for each non-selected object 6. Repeat steps 2, 3 and 4, until find the Medoids.
2.1.2 CLARA (Clustering large applications)-
Designed by Kaufman and Rousseeuw to handle large datasets, CLARA (Clustering large
Applications) relies on sampling. Instead of finding representative objects for the entire data set, CLARA draws
a sample of the data set, applies PAM on the sample, and finds the Medoids of the sample. The point is that, if
the sample is drawn in a sufficiently random way, the Medoids of the sample would approximate the Medoids of
the entire data set. To come up with better approximations, CLARA draws multiple samples and gives the best
clustering as the output. Here, for accuracy, the quality of a clustering is measured based on the average
dissimilarity of all objects in the entire data set, and not only of those objects in the samples.
CLARA Procedure-
1. Input the dataset D
2. Repeat n times
3. Draw sample S randomly from D
4. Call PAM from S to get Medoids M.
5. Classify the entire dataset D to Cost1.....cost k
6. Calculate the average dissimilarity from the obtained clusters
Complementary to PAM, CLARA performs satisfactorily for large data sets (e.g., 1,000 objects in 10 clusters).
2.1.3 CLARANS (A clustering algorithm based on randomized search)
It gives higher quality clusterings than CLARA, and CLARANS requires a very small number of searches. We
now present the details of Algorithm CLARANS.
Procedure of CLARANS-
1. Input parameters num local and max neighbour. Initialize i to 1, and min cost to a large number.
2. Set current to an arbitrary node in n:k.
3. Set j to 1.
4. Consider a random neighbour S of current, and based on 5, calculate the cost differential of the two nodes.
5. If S has a lower cost, set current to S, and go to Step 3.
6. Otherwise, increment j by 1. If j max neighbour, go to Step 4.
7. Otherwise, when j > max neighbour, compare the cost of current with min cost. If the former is less than
min cost, set min cost to the cost of current and set best node to current.
8. Increment i by 1. If i > num local, output best node and halt. Otherwise, go to Step 2.
Steps 3 to 6 above search for nodes with progressively lower costs. But, if the current node has already
been compared with the maximum number of the neighbors of the node (specified by max neighbor) and is still
of the lowest cost, the current node is declared to be a “local” minimum. Then, in Step 7, the cost of this local
minimum is compared with the lowest cost obtained so far. The lower of the two costs above is stored in min
cost. Algorithm CLARANS then repeats to search for other local minima, until num local of them have been
found.
As shown above, CLARANS has two parameters: the maximum number of neighbours examined (max
neighbour) and the number of local minima obtained (num local). The higher the value of max neighbour, the
closer is CLARANS to PAM, and the longer is each search of a local minima. But, the quality of such a local
minima is higher and fewer local minima need to be obtained.

III. Proposed Work
The procedure followed by partitioning algorithms can be stated as follows: “Given n objects, these
methods construct k partitions of the data, by assigning objects to groups, with each partition representing a
cluster. Generally, each cluster must contain at least one object; and each object may belong to one and only one
cluster, although this can be relaxed”. The present study analyzes the use of PAM , CLARA, CLARANS and
ECLARNS.
ENHANCED CLARANS (ECLARANS): This method is different from PAM, CLARA AND
CLARANS. Thus method is produced to improve the accuracy of outliers. ECLARANS is a partitioning
algorithm which is an improvement of CLARANS to form clusters with selecting proper nodes instead of
selecting as random searching operations. The algorithm is similar to CLARANS but these selected nodes
reduce the number of iterations of CLARANS ECLARANS Procedure. The Previous research established
ECLARANS as an effective algorithm for outlier detection but till now it doesn’t have better time complexity
thus by this research work we can also achieve this.
The algorithm is-
2. Calculating distance between each data points
3. Choose n maximum distance data points
4. Set current to an arbitrary node in n: k
5. Set j to 1.
6. Consider a random neighbour S of current, and based on 6, calculate the cost differential of the two nodes.
10. Increment i by 1. If i > num local, output best node and halt. Otherwise, go to Step 4.
Fig 2: Flowchart of ECLARANS algorithm
3.1 Proposed Methodology
In modified ECLARANS the approach of selecting nodes have been changed rather than selecting random
nodes after calculating the maximum cost between nodes we have chosen that points which are causing
maximum cost.
3.2 Modified Algorithm
2. Calculating distance between each data points for calculation select those points which has not been visited.
3. Select the maximum distance data points.
4. Set current to that node which is having highest distance if it is not been visited.
5. Set j to 1.

6. Consider a random neighbour S of current, and based on 6, calculate the cost differential
Between two nodes.
10. Increment i by 1. If i > num local, output best node and halt. Otherwise, go to Step4.
IV. Results
Proposed algorithm has been implemented using programming language Java and WHO dataset. The
java has been used using Netbeans7.3.1 which provides easy to implement graphical user interface for the
proposed system. Implemented software has been run using various lengths of data. Time required for various
executions has been recorded for different steps of the proposed work and results have been drawn. After
running various time the table of execution time (in seconds) are-
By the no. of execution analysis chart has been created which is a graphical presentation of results used
for comparison with other clustering algorithms
Fig 3: comparison of Different clustering algorithms for 8000 data objects

Fig 4: comparison of Different clustering algorithms for 8000 data objects
V. Conclusion
Modified ECLARANS has been found more accurate and time efficient. There are large number of
Partition based outlier detection technique are available. They can be used to solve all problems. But all
algorithms are designed under certain assumption and different algorithm is used under different condition. Such
as k-mean is used to handle spherical shaped cluster we cannot used to find arbitrary shaped cluster. The main
aim of this clustering algorithm is , outlier detection with improved time efficiency and outlier detection
accuracy. Additionally, the efficiency and effectiveness of a novel outlier detection algorithm can be defined as
to handle large volume of data as well as high-dimensional features with acceptable time and storage, to detect
outliers in different density regions, to show good data visualization and provide users with results that can
simplify further analysis.
References
[1]. A. Mira, D.K. Bhattacharyya, S. Saharia,” RODHA: Robust Outlier Detection using Hybrid Approach”, American Journal of
Intelligent Systems, volume 2, pp 129-140, 2012
[2]. Al-Zoubi M. “An Effective Clustering-Based Approach for Outlier Detection”(2009)
[3]. A K Jain,M N Murthy. “Data Clustering A Review” ACN Computing Surveys Vol 31,No3.September 1999.
[4]. D Moh, Belal Al-Zoubi, Ali Al-Dahoud, Abdelfatah A Yahya “New outlier detection method based on fuzzy clustering”2011.
[5]. Deepak Soni, Naveen Jha, Deepak Sinwar,” Discovery of Outlier from Database using different Clustering Algorithms”, Indian J.
Edu. Inf. Manage., Volume 1, pp 388-391, September 2012.
[6]. Han & Kamber & Pei,” Data Mining: Concepts and Techniques (3rded.) Chapter 12, ISBN-9780123814791
[7]. Ji Zhang,” Advancements of Outlier Detection: A Survey”, ICST Transactions on Scalable Information Systems, Volume 13, pp 1-
26 January-March 2013
[8]. Maria Halkidi, Yannis Batistakis, Michalis Vazirgiannis,” On Clustering Validation Techniques”, Journal of Intelligent Information
Systems, pp 107–145, January 2001.
[9]. Maria Kontaki, Anastasios Gounaris, Apostolos N. Papadopoulos, Kostas Tsichlas, Yannis Manolopoulos,” Continuous Monitoring
of Distance-Based Outliers over Data Streams”, Proceedings of the 27th IEEE International Conference on Data Engineering ,
Hannover, Germany, 2011.
[10]. Moh'd belal al-zoubi1, ali al-dahoud2, abdelfatah a. yahya” New Outlier Detection Method Based on Fuzzy Clustering”
[11]. Mr Ilango, Dr V Mohan,” A Survey of Grid Based Clustering Algorithms”, International Journal of Engineering Science and
Technology, Volume 2, pp 3441-3446, 2010.
[12]. Ms. S. D. Pachgade, Ms. S. S. Dhande,” Outlier Detection over Data Set Using Cluster-Based and Distance-Based Approach”,
International Journal of Advanced Research in Computer Science and Software Engineering, Volume 2, pp 12-16 June 2012.
[13]. Periklis Andritsos,” Data Clustering Techniques”, pp 1-34, March 11, 2002.
[14]. P. Murugavel, Dr. M. Punithavalli,” Improved Hybrid Clustering and Distance-based Technique for Outlier Removal”, International
Journal on Computer Science and Engineering, Volume 3, pp 333-339, 1 January 2011.
[15]. Sivaram, Saveetha,”AN Effective Algorithm for Outlier Detection”, Global Journal of Advanced Engineering Technologies,
Volume 2, pp 35-40, January 2013.
[16]. S.Vijayarani, S.Nithya,” Sensitive Outlier Protection in Privacy Preserving Data Mining”, International Journal of Computer
Applications, Volume 33, pp 19-27, November 2011.
[17]. S.Vijayarani, S.Nithya,” An Efficient Clustering Algorithm for Outlier Detection”, International Journal of Computer Applications,
Volume 32, pp 22-27, October 2011
[18]. Silvia Cateni, Valentina Colla ,Marco Vannucci Scuola Superiore Sant Anna, Pisa,” Outlier Detection Methods for Industrial
Applications”, ISBN 78-953-7619-16-9, pp. 472, October 2008
[19]. Shalini S Singh, N C Chauhan,” K-means v/s K-medoids: A Comparative Study”, National Conference on Recent Trends in
Engineering & Technology, May 2011.
[20]. Tan, Steinbach, Kumar,” Introduction to Data Mining (1sted.) chapter 10”, ISBN-0321321367.

Performance Analysis of Different Clustering Algorithm

Recommended

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Performance Analysis of Different Clustering Algorithm (20)

More from IOSR Journals (20)

Recently uploaded (20)

Performance Analysis of Different Clustering Algorithm