PRIVACY PRESERVING DATA MINING BASED ON VECTOR QUANTIZATION

International Journal of Database Management Systems ( IJDMS ) Vol.5, No.2, April 2013
DOI : 10.5121/ijdms.2013.5203 29
PRIVACY PRESERVING DATA MINING BASED
ON VECTOR QUANTIZATION
D.Aruna Kumari1
, Dr.K.Rajasekhara Rao2
, M.Suman3
1,3
Department of Electronics and Computer Engineering, Associate professors ,CSI Life
Member K.L.University, Vaddeswaram,Guntur
1
Aruna_D@kluniversity.in and 3
suman.maloji@gmail.com
2
Department of Computer Science and Engineering ,professor, K.L.University,
Vaddeswaram,Guntur
2
krr@kluniversity.in
1,2,3
CSI LIFE MEMEBERS, 3
CSI-AP Student CO-coordinator
ABSTRACT
Huge Volumes of detailed personal data is continuously collected and analyzed by different types of
applications using data mining, analysing such data is beneficial to the application users. It is an important
asset to application users like business organizations, governments for taking effective decisions. But
analysing such data opens treats to privacy if not done properly. This work aims to reveal the information
by protecting sensitive data. Various methods including Randomization, k-anonymity and data hiding have
been suggested for the same. In this work, a novel technique is suggested that makes use of LBG design
algorithm to preserve the privacy of data along with compression of data. Quantization will be performed
on training data it will produce transformed data set. It provides individual privacy while allowing
extraction of useful knowledge from data, Hence privacy is preserved. Distortion measures are used to
analyze the accuracy of transformed data.
KEYWORDS
Vector quantization, code book generation, privacy preserving data mining ,k-means clustering.
1. INTRODUCTION
Privacy preserving data mining (PPDM) is one of the important area of data mining that aims to
provide security for secret information from unsolicited or unsanctioned disclosure. Data mining
techniques analyzes and predicts useful information. Analyzing such data may opens treat to
privacy .The concept of privacy preserving data mining is primarily concerned with protecting
secret data against unsolicited access. It is important because Now a days Treat to privacy is
becoming real since data mining techniques are able to predict high sensitive knowledge from
huge volumes of data[1].
Authors Agrawal & Srikant introduced the problem of “privacy preserving data mining” and it
was also introduced by Lindell & Pinkas. Those papers have concentrated on privacy preserving
data mining using randomization and cryptographic techniques. Lindell and Pinkas designed new
approach to PPDM using Cryptography but cryptography solution does not provides expected

30
accuracy after mining result. And agrawal and srikanth focused on randomization and preserving
privacy when the data is taken from multiple parties. When the data is coming from multiple
sources then also privacy should be maintained. Now a days this privacy preserving data mining
is becoming one of the focusing area because data mining predicts more valuable information that
may be beneficial to the business, education systems, medical field, political ,…etc.
2. RELATED WORK
Many Data modification techniques are discussed in [1 ][3][4]
A. Perturbation or Randomization:
Agrawal and Srikant (2000) Introduced the randomization algorithm for PPDM, Randomization
allows a several number of users to submit their sensitive data for effective centralized data
mining while limiting the disclosure of sensitive values. It is relatively easy and effective
technique for protecting sensitive electronic data from unauthorized use. In this case there is one
server and multiple clients will operate ,Clients are supposed to send their data to server for
mining purpose , in this approach each client adds some random noise before sending it to the
server. So Sever will perform mining on that randomized data.
B. Suppression
Another way of preserving the privacy is suppressing the sensitive data before any disclosure or
before actual mining takes place. Generally Data contains several attributes, where some of the
attributes may poses personal information and some of the attributes predicts valuable
information. So we can suppress the attributes in particular fashion that reveals the personal
information.
They are different types are there
1. Rounding
2. Generalization
In rounding process the values like 23.56 will be rounded to 23 and 25.77 rounded to
26,…etc
In generalization process, values will be generalized like an address is represented with zip
code.
if data mining requires full access to the entire database at that time all this privacy
preserving data mining techniques are not required.
C. Cryptography
This is Also one of the famous approach for data modification techniques, Here Original Data
will be encrypted and encrypted data will be given to data miners. If data owners require original
data back they will apply decryption techniques.
3. VECTOR QUANTIZATION
This is the new technique proposed by (D.Aruna Kumari, Dr.k.Rajasekhara Rao, Suman,)), it
transforms the original data to a new form using LBG. The design of a Vector Quantization
consist of following steps:

• Design a codebook from input training data set;
• Encoding the original point of data with the indices of the nearest code vectors in the
codebook;
• Use index representation so as to reconstruct the data by looking up in the codebook.
For our PPDM problem , reconstructing the original data is not required,
involved such that it is difficult to get the original data back hence privacy is preserved.
LBG Design Algorithm
1. Given . Fixed
2. Let and
Calculate
3. Splitting: For
Set .
4. Iteration: Let
i. For
over all . Let
ii. For
iii. Set .
iv. Calculate
Design a codebook from input training data set;
ding the original point of data with the indices of the nearest code vectors in the
Use index representation so as to reconstruct the data by looking up in the codebook.
For our PPDM problem , reconstructing the original data is not required, so above two steps are
to be a ``small'' number.
, set
. Set the iteration index .
, find the minimum value of
. Let be the index which achieves the minimum. Set
, update the codevector
.
31
ding the original point of data with the indices of the nearest code vectors in the
Use index representation so as to reconstruct the data by looking up in the codebook.
so above two steps are
be the index which achieves the minimum. Set

v. If
vi. Set
as the final codevectors.
5. Repeat Steps 3 and 4 until the desired number of codevectors is obtained.
codebook is generated, one can perform transformation using quantization
Results :
We have implemented above LBG algorithm using Matlab
output screen shots Blue line represents original data and red line represents Codebook that is
compressed form of original data , hence it does not reveal the complete original information and
it will reveal only cluster centroids
Figure 1: VQ based on LBG design Figure 2: VQ based on LBG design Algorithm y=x;
Algorithm y=exp(x);
Figure 3: VQ based on LBG design Algorithm
y=x2; y=sin(x);
, go back to Step (i).
. For , set
as the final codevectors.
Repeat Steps 3 and 4 until the desired number of codevectors is obtained.
codebook is generated, one can perform transformation using quantization
We have implemented above LBG algorithm using Matlab Software, and tested the results. In the
luster centroids.
Figure 3: VQ based on LBG design Algorithm Figure 4: VQ based on LBG design Algorithm
32
Repeat Steps 3 and 4 until the desired number of codevectors is obtained. Once the
Software, and tested the results. In the
Algorithm

33
Backup And recovery of Information Loss
Generally Privacy preserving data mining , we apply some techniques for modifying data and
that modified data will be given to data miners.
In this paper we also concentrates keeping of original data as it is, so whenever data miners or
owners of that data requires original data , they will get it by maintaining a backup copy of that
data.
1. Taking all the data in to a DB1
2. We have to copy all the data in DB1 to DB2 (DB1=DB2)
3. Perform Data modification using LBG Design Algorithm and store the data in to DB1
4. Write the program to delete the all the contents in the backup of DB1,(so no except the
users cannot access the information in DB1)
5. To see the actual details and to modify the details the admin should access the
information in DB2
6. So, far the users know only DB1, data base owners can access to DB2
Bit Error Rate
In Data transmission, the number of bit errors is the number of received bits of a data stream
over a communication channel that have been altered due to noise, interference, distortion or bit
synchronization errors.
In our problem, we are transforming the original data to some other form using vector
quantization. Hence we need to calculate the bit error rate for compressed data.
Always we try to minimise the bit error rate for accuracy
For example
Original data is
1 0 1 0 1 1 0 0 1 0
And after the transformation, the received data is
1 1 1 0 1 1 0 1 1 0
(Two errors are there, i.e, we are not receiving exact data only 80% accuracy is achieved because
of two bit errors)
4. CONCLUSIONS
This work is based on vector quantization , it is a new approach for privacy preserving data
mining, upon applying this encoding procedure one cannot reveal the original data hence privacy
is preserved. At the same time one can get the accurate clustering results. Finally we would like
conclude that Efficiency depends on the code book generation.

34
REFERENCES
[1] D.Aruna Kumari , Dr.K.Rajasekhar rao, M.suman “ Privacy preserving distributed data mining using
steganography “In Procc. Of CNSA-2010, Springer Libyary
[2] T.Anuradha, suman M,Aruna Kumari D “Data obscuration in privacy preserving data mining in Procc
International conference on web sciences ICWS 2009.
[3] Agrawal, R. & Srikant, R. (2000). Privacy Preserving Data Mining. In Proc. of ACM SIGMOD
Conference on Management of Data (SIGMOD’00), Dallas, TX.
[4] Alexandre Evfimievski, Tyrone Grandison Privacy Preserving Data Mining. IBM Almaden Research
Center 650 Harry Road, San Jose, California 95120, USA
[5] Agarwal Charu C., Yu Philip S., Privacy Preserving Data Mining: Models and Algorithms, New
York, Springer, 2008.
[6] Oliveira S.R.M, Zaiane Osmar R., A Privacy-Preserving Clustering Approach Toward Secure and
Effective Data Analysis for Business Collaboration, In Proceedings of the International Workshop on
Privacy and Security Aspects of Data Mining in conjunction with ICDM 2004, Brighton, UK,
November 2004.
[7] Wang Qiang , Megalooikonomou, Vasileios, A dimensionality reduction technique for efficient time
series similarity analysis, Inf. Syst. 33, 1 (Mar.2008), 115- 132.
[8] UCI Repository of machine learning databases, University of California,
Irvine.http://archive.ics.uci.edu/ml/
[9] Wikipedia. Data mining. http://en.wikipedia.org/wiki/Data_mining
[10] Kurt Thearling, Information about data mining and analytic technologies http://www.thearling.com/
[11] Flavius L. Gorgônio and José Alfredo F. Costa “Privacy-Preserving Clustering on Distributed
Databases: A Review and Some Contributions
[12] D.Aruna Kumari, Dr.K.rajasekhar rao,M.Suman “Privacy preserving distributed data mining: a new
approach for detecting network traffic using steganography” in international journal of systems and
technology(IJST) june 2011.
[13] Binit kumar Sinha “Privacy preserving clustering in data mining”.
[14] C. W. Tsai, C. Y. Lee, M. C. Chiang, and C. S. Yang, A Fast VQ Codebook Generation Algorithm
via Pattern Reduction, Pattern Recognition Letters, vol. 30, pp. 653{660, 2009}
[15] K.Somasundaram, S.Vimala, “A Novel Codebook Initialization Technique for Generalized Lloyd
Algorithm using Cluster Density”, International Journal on Computer Science and Engineering, Vol.
2, No. 5, pp. 1807-1809, 2010.
[16] K.Somasundaram, S.Vimala, “Codebook Generation for Vector Quantization with Edge Features”,
CiiT International Journal of Digital Image Processing, Vol. 2, No.7, pp. 194-198, 2010.
[17] Vassilios S. Verykios, Elisa Bertino, Igor Nai Fovino State-of-the-art in Privacy Preserving Data
Mining in SIGMOD Record, Vol. 33, No. 1, March 2004.
[18] Maloji Suman,Habibulla Khan,M. Madhavi Latha,D. Aruna Kumari “Speech Enhancement and
Recognition of Compressed Speech Signal in Noisy Reverberant Conditions “ Springer -Advances in
Intelligent and Soft Computing (AISC) Volume 132, 2012, pp 379-386

PRIVACY PRESERVING DATA MINING BASED ON VECTOR QUANTIZATION

More Related Content

What's hot (20)

Similar to PRIVACY PRESERVING DATA MINING BASED ON VECTOR QUANTIZATION (20)

Recently uploaded (20)

PRIVACY PRESERVING DATA MINING BASED ON VECTOR QUANTIZATION