International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 8August 2013

ISSN: 2231-2803 http://www.ijcttjournal.org Page 2731

Semisupervised Learning Taxonomy-Aware
Catalog Integration
D.Umavathi M.Sc
1
, R.Tamil selvi

M.Sc., M.Phil
2
,
1
Research Scholar, Department of Computer Science, Dr. SNS Rajalakshmi College of Arts and Science, Coimbatore,
Tamil Nadu
2
Assistant Professor, Department of Computer Science, Dr. SNS Rajalakshmi College of Arts and Science, Coimbatore,
Tamil Nadu


Abstract Data integration is the major important task for
online commercial portals and commerce search engine based
applications. The data integration task faced by online
commercial portals and e-commerce search engines are the
integration of products coming from multiple providers to their
product catalogs. Categorization of products from the data
provides into the master taxonomy and whereas make use of the
data provider taxonomy information becomes major problem.
Overcome this problem we classify the products based on their
textual based classifier and taxonomy-aware step that adjusts the
results of a textual based classifier to ensure that products that
are close together in the provider taxonomy remain close in the
master taxonomy. In taxonomy aware calibration step the base
classifier derives the parameters value by tuning the values.
Existing base classifier step, still it becomes major problem
identify candidate products for labeling the products, to
overcome these problem we proposed a Semi supervised learning
techniques to incrementally retrain the base classifier with
elements chosen during the taxonomy-aware calibration step.
Proposed system it categorizes the product based on their
parameters choosen from calibration. Semi supervised learning
algorithm involves a large amount of unlabeled product data
with only a small number of labeled product data.
Semisupervised based active learning method identifies the
candidate products for labeling. Proposed system finds the each
candidate parameter 0
|
and then find the optimal parameter
y such that the improve the accuracy on the validation set is
maximized. An experimental result shows that the Semi
supervised learning algorithm that are efficient and thus
applicable to the large data sets that are typical on the web.

Keywords Catalog integration, classification, data mining,
taxonomies, Semi supervised learning.
I. INTRODUCTION
In online shopping and e commerce based applications,
ever-increasing the numeral of web portals present a user
experience about the shopping. This web portals includes
several commercial sites such as Amazon and Shopping.com
and commerce search engines such as Google Product Search
and Bing Shopping.
For this purpose the data integration task is important for
many web portals and commercial portals .The data
integration task is faced by these marketable portals is the
mixing of data upcoming fromnumerous data providers into a
particular product catalog. It is known as is product
categorization. All of the web portals maintain the own master
taxonomy for organizing products and it is used for both
browsing and searching purposes. When a new products arrive
fromthe dissimilar providers, it automatically categorize the
products in master taxonomy according to their users. But in
website environment it becomes difficult to assign the
products fromtheir catalog to the appropriate category in the
master taxonomy. To do this process we need automatic
labeling techniques for categorization products coming from
data providers.
Automatically categorization of the product scenario each
data providers includes their own taxonomy and their products
are previously associated with a provider taxonomy category.
It will differ fromthe normal master taxonomy model. In this
scenario product taxonomy model the products that are in
nearby categories should be classified into nearby categories
in the master taxonomy. In this paper first propose the
technique that forces to solve the above mentioned problems.
First use the taxonomy information to adjust the results of a
text-based classifier. The text based classifier model makes
use of the taxonomy structure of the master and provider
taxonomies in order to achieve relationships among the
dissimilar categories in the taxonomy with the purpose of
relying on the category membership information of the
products. In the text based classifier new products arrives in
the web portals or ecommerce applications information fails to
achieve the clear classification or labeling of the class label
with supervised learning and the category-membership
information of the product creation in the provider taxonomy,
because it does not provide any further help out, while there is
no obvious mapping among categories at the leaf level. To
overcome these problems we proposed a semi supervised
based classifier to incrementally retrain the base classifier
with elements chosen during the taxonomy-aware calibration
step. Proposed systemmethods construction uses jointly of
labeled and unlabeled data are called semi-supervised learning
at taxonomy calibration step.

The major steps of the work as follows:
1. First originate the taxonomy-aware catalog integration
difficulty as a structured prediction difficulty. In this method
the approach that leverages the structure of the taxonomies in
order to enhance catalog integration.
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 8August 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2732


2. Second describe the taxonomy aware classification
process with two steps: In first step product are classified
under base classification step, then use taxonomies aware
processing steps.
3. During the taxonomy aware classification step the
optimization problem or label classification problem have
been overcomed with scalable algorithmfor the taxonomy-
aware classification process.
4.Tuning the parameters of the k, 0,y is important for the
performance of the system .Semi supervised learning
algorithm makes best result for classification
results .Proposed systemwe apply Semi supervised learning
algorithmfor selection best base classifier b on the products of
the validation set.
5. Finally evaluate the experimental results on real-world
data and compare taxonomy- aware classification, proposed
semi supervised learning parameter calibration provides a
significant improvement in accuracy over existing state-of-
the-art classifiers.
.
II. RELATED WORK
In this section study the various catalog integration
problem and methods followed to solve the catalog integration
problemas well as Metric labeling and structured prediction.
R. Agrawal and R. Srikant [1] introduce the problemof
pervasive in web portal environment. It automatically
processes the product catalog to build the base classifier for
product integration of documents in the master catalog for
predicting the category of unknown documents. Our solution
insight is that numerous of the data sources have their have
possession of categorization and classification accuracy can
be improved by factoring in the implicit information in these
source categorizations. It makes use of source category
information, but treats the basis and target taxonomy as flat.
Sarawagi et al [2] establish cross training model with semi-
supervised learning for document classification occurrence of
multiple label sets. Document classification is a well-
established region of text mining. A document classifier is
original trained using documents with preassigned labels or
classes picked froma set of labels it is named as taxonomy or
catalog. Once the classifier is trained, it is offered test
documents for which it must guess the best labels.
Zhang and Lee have also developed approaches to catalog
integration by using boosting [4] and transductive learning [3],
[5]. Even though these approaches attain better categorization
accuracy similar to the cross-training approach, they
necessitate training data that are labeled in both the source and
the target taxonomies. Consequently such method is not
appropriate to our problemsetting.
Matching problemfor categorization of the products into
the taxonomy model ontology matching and alignment
schema was proposed in previous work. Glue [7] introduces a
machine learning method to learn how to map the product
between the products with ontology matching and alignment
of the product. Iliads [8] introduces machine learning and
logical inference technique to production alignments. In
general, the focus in ontology alignment is to map nodes of
source taxonomy to nodes of target taxonomy. In dissimilarity
metrics the similarity of the systemnot interested in solving
the (much harder) alignment problem between taxonomies,
but rather given an instance (i.e., a product) the goal is to
categorize it in the target taxonomy using aids from the
taxonomy structure. The end objective is forever the
classification of the product. This dissimilarity is very vital in
many realistic scenarios.
Nandi and Bernstein [9] recommend an approach for
corresponding taxonomies based on query term distributions.
Primary it perform the mapping at the taxonomy level,
mapping category from the source to the target, while we
achieve the mapping at the occurrence level by categorizing
personality product instances to the target taxonomy.
Following the approach is not based on classification but
rather on exploiting distributions of terms associated with the
categories.
Formulation of the catalog integration problem as an
optimization problem is stimulated by the metric labeling
problemthat was introduced in [10]. In the metric labeling
problem is to discover the optimal labeling of a number of
objects consequently that they reduce an assignment and a
separation cost. The problemis NP-hard [10] and the different
obtainable estimated solutions formulate it as an LP [10] or a
QP [11]. The difficulty of all these method makes them
unsuitable to large-scale data sets with additional than a few
hundreds of products. The purpose of our optimization
problem is also comparable to the objective that arises in
computer vision problems [12], [13], [14].
The most popular submission is image restoration;
anywhere the goal is to re-establish the intensity of each pixel
in a picture by means of the values of the experiential
intensities. The algorithms developed in this region focus on
division of costs definite in the euclidean space, i.e., the
relationship of two items decrease linearly through their
euclidean distance. Even though such algorithms are scalable,
they cannot be personalized to the separation cost definition
that is appropriate for taxonomies.

III. TAXONOMY-AWARE CATALOG INTEGRATION AND SEMI
SUPERVISED LEARNING ALGORITHM FOR PARAMETER
CALIBRATION
First we formulate the taxonomy-catalog integration
problem to establish a some basic terminology. A creation of
product x is an itemthat can be buy at a commercial portal.
Every product has a textual demonstration that consists of a
name and possibly a set of attribute-value pairs. In this step
the product taxonomy can be represented as a Graph G={V, E}
with a directed acyclic graph (DAG) whose nodes Cg
represent the set of probable categories into which products
are organized. Each graph in an edge ( c
1
,c
2
) Eg
represents a subsumption association between two categories
c
1
and c
2
. After defining of the taxonomy integration
problem then describe our move toward to taxonomy aware
categorization as a two-step process. Primary each product is
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 8August 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2733


classifying use a base classifier without aware of the
taxonomies. Afterward use the formation of the source and
target taxonomies in regulate to correct the output of the base
classifier and construct a final classification. It is named as the
taxonomy-aware processing step.
These steps are described below:
A. The Base Classification Step
The In the base foundation classification step categorize the
products based on their textual demonstration. In this
classification step train the text based classifier using machine
learning methods such as naive bayes (NB) and Logistic
Regression (LR) where separation of the target catalog as the
training set. It provides us with examples of products label
with category of the target taxonomy. The features of the
classifier are extended fromthe textual product demonstration.
Let b indicate the classification representation after training
process competition.
B. The Taxonomy-Aware Processing Step
After that the taxonomy-aware processing step is to
categorize the target taxonomy results from the foundation
classification step by taking into account the associations of
the products in the source and target taxonomies. In this step
doesnt handle all the products in efficient manner,
optimization problemoccurs, to overcome these converse the
different parameters of the problem. It is defined fromthe
given a source catalog K
s
, and a target catalog K
t
, the
objective is to find a labeling vector that minimizes the
following cost function:
C0SI(
S
,
t
,) =
(1) A Cost(x,
x
) +
xcP
s
S Cost(x,y,
x
,

)
x,cP
s

The taxonomy-aware procedure f
T
is the algorithm that
finds the labeling that minimizes the cost function:

1
(
S
,
t
) =argmin

C0SI(
S
,
t
,)
To classify the products fromthe base classifier calculate
probabilities of the base classifier to define the task of cost
function. A COST: P
s
C
t

+
.For a product x the cost of
classifying product x to objective category
x
is defined as
follows:

A Cost(x,
x
) =1Pr
b
(
x
|

)

Important similarity description is supposed to assure the
perception those two categories that are close up in the
taxonomy tree are more comparable than two categories that
are far separately. For example, two categories that have a
general parent are further similar than two categories that have
dissimilar parents and an ordinary grandparent. The division
cost as a function of the similarity sim
S
(s
x
,s

) between
categories and of x and y in the source taxonomy S and the
similarity
S Cost(x,y,
x
,

) =o(sim
S
(s
x
,s

),sim
1
(s
x
,s

))
Optimization problemhave been occurs in all of the above
mentioned steps, to overcome these problems, scalable
algorithmfor the taxonomy-aware categorization step to large
data sets. Even though present our method with respect to our
exact problem. It can be applied to other prearranged
prediction problems in arrange to deal with the quadratic
numeral of pairwise relationships. To performthis process
using search pruning methods and then proceed calibration
step to categorize the master and product taxonomy.

Search Space Pruning presents a heuristic for proficiently
performing arts the taxonomy-aware calibration step. The idea
is to thoughtfully fix the group or category for a number of
products in the foundation catalog in order to achieve a
landscape of the mappings among the two taxonomies. From
this define the subset of products that categorize the
products .Let 0e [0,1] be a threshold value that define while
the category probability approximation returned by the base
classifier is great enough therefore that the predicted category
is expected to be accurate. Let F
0
be the subset of products
that pass the threshold is defined as,
F
0
={x P
s
| max
ycC
t
Pr
b
[|x 0}

x
=argmax
ycC
t
Pr
b
[|x]
Let 0
0
=P
s
/ F
0
denote the products whose classification
remains open. Each open product 0
0
autonomously and
calculate a division cost for only with respect to the fixed
products in F
0
.If s
x
is the source category of x and t
x
is a
candidate target category, then the separation cost for this
source-target pair is defined as follow:
(s
x
,t
x
)
= S C0SI(S Cost(s
x
,o,t
x
,)n
ccS,:c1
(s
x
,t
x
)n(

o,)
Algorithm: TACI algorithm
Input: Source catalog
S
, Target Taxonomy T, base
classifier b and parameters 0,k,y
Output: Labeling vector
1. F
s
0
2. For all x P
s
do
3.

argmox
:cC
t
,max
ycC
t
Pr
b
[|x]
4. if Pr
b
[

|x] 0 then
5.
x

6. F
0
F
0
{x}
7. Else
8. 0
0
0
0
{x}
9. Compute TOP
k
(x)
10. Compute candidate pairs E
0,k

11. Initialize hash table EI to empty
12. For all (o,)eE
0,k
do
13. HT(o,)=H(o,))
14. For all x 0
0
do
15.
x
argmin
:cTOPk (x)
{(1y)A C0SI
x,:+yH1(S
x
,:)
}
.


C. Parameter calibration
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 8August 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2734


The tuning of the parameters k, 0, and y is important for
the performance of our algorithm. The validation set consists
of products that are cross labeled in both the source and the
target taxonomy. Base classifier training that involves tens of
millions of features, while it is big enough to tune few
parameters of the TACI algorithm. The first parameter set is
parameter k, such that the accuracy of the classifier over the
top-k categories is high. The details are described below. Then,
we tune the parameters 0, and. For each candidate parameter
we find the optimal parameter y such that the accuracy of
the TACI algorithmon the validation set is maximized. Notify
all the parameters that are selected such as to maximize the
accuracy of the TACI algorithmon the validation set.
D. Semi supervised learning for calibration step
In generally the learning methods can be divided into
supervised and unsupervised learning methods. The
supervised learning methods learner aims at estimation of the
input output relationship by using objective function with
training set data set {x
i
, y
i
}, i =1, . . . , N where the inputs x
are n-dimensional vectors and the labels y are continuous
values for regression tasks and discrete for classification
problems; In unsupervised learning only the raw data x
i
are
available, not including the consequent labels y
i
. This type of
the algorithm belonging to the group are clustering and
independent component analysis routines .It becomes difficult
to handle the unlabeled data, to handle this situation where
some labeled patterns are provided jointly with unlabeled ones
arise frequently. This type of learning is named as the semi
supervised learning. Proposed algorithmfor semi-supervised
learning during calibration step that on one hand is easy to
execute and on the other hand is guaranteed to improve the
categorization of the product result performance.
The main idea of the proposed algorithmis to estimate the
top eigen functions of the integral operator from the both
labeled and unlabeled examples, and learn from the labeled
examples the best prediction function in the subspace spanned
by the estimated eigen functions. Let X is a compact domain
or a manifold in the Euclidean space R
d.
Let D ={x
i
, i =
1, . . . , N x
i
X} be a collection of training examples.
Randomly select the n examples fromD for labeling. Without
loss of generality, we assume that the first n examples are
labeled by y
l
=(y
1
, . . . , y
n
)

R
n
. We denote by y =(y
1
, . . . ,
y
N
)

R
N
the true labels values for parameters such as for all
the examples in D. In this study, we assume y = f(x) is
decided by an unknown deterministic function f(x). Our goal
is to learn an accurate prediction parameter 0 to
incrementally retrain the base classifier at calibration step.
Algorithm 2: Semisupervised learning for calibration step
Input: D ={x
i
, i =1, . . . , N x
i
X} be a collection of
training examples, y
l
=(y
1
, . . . , y
n
)

labels for the n
examples selected randomly, s be the eigen vectors selected
1. Compute (
i

,z
i

),i=1,,s the first eigen functions and

eigen values for the integral operator is defined as
I
N

()(.) =
1
N
k(x

N
=1
,.).(x

)
2.Compute the prediciton result g(.) such as considered
prediction parameter 0 to incrementally retrain the base
classifier at calibration step g (X) = y
]
s
]=1

]

(X) Where
y

={y
1

,,y
s

} is given by solving the following equation,

y

=argmin
ycR
s (y
]

(X

) y

s
]=1
n
=1
)
2
.
IV. EXPERIMENTAL RESULTS
Finally in this section measure the classification accuracy
of the taxonomy classification step for TACI algorithmand
TACI with semi supervised learning at calibration step. The
results show the benefits of our taxonomy-aware calibration
step and compare the taxonomy-aware algorithm, semi
supervised calibration step. Measure the accuracy for all
methods by considering the three different providers such as
Amazon, Etilize, and Pricegrabber. The use as master catalog
the catalog of Bing Shopping, which aggregate data feeds
from retailers, distributors, resellers, and other profitable
portals. In all the experiments, consider a target taxonomy that
consists of all the categories in Bing Shopping taxonomy that
is related to consumer electronics. Taxonomy-Aware Catalog
Integration with Naive bayes (TACI-NB), Taxonomy-Aware
Catalog Integration with Linear regression (TACI-LR).
A. Classification Accuracy Evaluation
In this section, we compare the classification accuracy of
the different approaches for catalog integration. The results
for all algorithms over all data sets are in Table 1 and the
corresponding figure are shown in Figure 1.
Providers TACI-
NB
TACI-
LR
TACI-
SSL Amazon 81.1 75.9 84.7
Etilize 81.7 91.8 93
pricegrab
ber
71.2 74.4 83.5
. Table1: Classification Accuracy Evaluation

Figure 1: Classification Accuracy Evaluation



B. Time Comparison Accuracy Evaluation
In this section, we compare the time comparison accuracy
of the different approaches for catalog integration. The results
for all algorithms over all data sets are in Table 2 and the
corresponding figure are shown in Figure 2.
Providers TACI-
NB
TACI-
LR
TACI-
SSL
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 8August 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2735


Amazon 775 682 600
Etilize 75 50 60
Pricegrabb
er
350 300 230
Table 2: Time comparison Accuracy Evaluation.


Figure 2: Time comparison Accuracy Evaluation
IV. CONCLUSION
In this research we presented a well-organized approach to
catalog integration that is based on the use of source category
and taxonomy structure information. The proposed semi
supervised learning algorithmwere used for retrain the base
classifier during the product calibration step, they can also be
used for other problems. The output of the parameter result as
chosen might be second-hand as a feature for itemidentical,
while would like to match elements classified under the
master taxonomy to incoming offers from the providers.
Experimental results also showed that this move toward leads
to considerable gains in correctness with respect than the
existing calibration step based classifier.
REFERENCES
[1] R. Agrawal and R. Srikant, On Integrating Catalogs, Proc. 10
th
Intl
Conf. World Wide Web (WWW), pp. 603-612, 2001.

[2] S. Sarawagi, S. Chakrabarti, and S. Godbole, Cross-Training:
Learning Probabilistic Mappings between Topics, Proc. Ninth ACM
SIGKDD Intl Conf. Knowledge Discovery and Datamining (KDD),
2003.

[3] D. Zhang and W.S. Lee, Web Taxonomy Integration through Co-
Bootstrapping, Proc. 27th Ann. Intl ACM SIGIR Conf. Research and
Development in Information Retrieval, pp. 410-417, 2004.

[4] D. Zhang and W.S. Lee, Web Taxonomy Integration Using Support
Vector Machines, Proc. 13th Intl Conf. World Wide Web (WWW),
pp. 472-481, 2004.

[5] D. Zhang, X. Wang, and Y. Dong, Web Taxonomy Integration Using
Spectral Graph Transducer, Proc. ER Workshop, pp. 300- 312, 2004.

[6] A. Nandi and P.A. Bernstein, Hamster: Using Search Clicklogs for
Schema and Taxonomy Matching, Proc. VLDB Endowment, vol. 2,
no. 1, pp. 181-192, 2009.

[7] A. Doan, J . Madhavan, R. Dhamankar, P. Domingos, and A. Halevy,
Learning to Match Ontologies on the Semantic Web, The VLDB J .,
vol. 12, no. 4, pp. 303-319, 2003.

[8] O. Udrea, L. Getoor, and R.J. Miller, Leveraging Data and Structure
in Ontology Integration, Proc. ACM SIGMOD Intl Conf.
Management of Data, pp. 449-460, 2007.

[9] A. Nandi and P.A. Bernstein, Hamster: Using Search Clicklogs for
Schema and Taxonomy Matching, Proc. VLDB Endowment, vol. 2,
no. 1, pp. 181-192, 2009.

[10] J . Kleinberg and E. Tardos, Approximation Algorithms for
Classification Problems with Pairwise Relationships: Metric Labeling
and Markov RandomFields, J. ACM, vol. 49, no. 5, pp. 616-639,
2002.

[11] P. Ravikumar and J . Lafferty, Quadratic Programming Relaxations
for Metric Labeling and Markov RandomField Map Estimation, Proc.
23rd Intl Conf. Machine Learning (ICML), pp. 737-744, 2006.

[12] Y. Boykov and V. Kolmogorov, An Experimental Comparison of Min-
Cut/Max-Flow Algorithms for Energy Minimization in Vision, IEEE
Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 9, pp.
1124-1137, Sept. 2004.

[13] Y. Boykov, O. Veksler, and R. Zabih, Fast Approximate Energy
Minimization via Graph Cuts, IEEE Trans. Pattern Analysis Machine
Intelligence, vol. 23, no. 11, pp. 1222-1239, Nov. 2001.

[14] V. Kolmogorov and R. Zabih, What Energy Functions can be
minimized via Graph Cuts? IEEE Trans. Pattern Analysis and
Machine Intelligence, vol. 26, no. 2, pp. 147-159, Feb. 2004