International Journal of Computer Techniques - Volume 3 Issue 2, Mar-Apr 2016

RESEARCH ARTICLE

OPEN ACCESS

Machine Learning Spectrum for

Web Data Analytics
1
Sunny Sharma

Department of Computer Science, Guru Nanak Dev University, Amritsar, Punjab

----------------------------------------************************----------------------------------

Abstract:
There are numerous ways to analyse the web information, generally web substance are housed in
large information sets and basic inquiries are utilized to parse such information sets. As the requests
expanded with time, mining web information amended to meet challenging task in a web analysis.
Machine learning methodologies are the most up to date one to go into these analysis forms. Different
approaches like decision trees, association rules, Meta heuristic and basic learning methods are embraced
for making web data appraisal and mining data from various web instances. This study will highlight these
approaches in perspective of web investigation. One of the prime goals of this exploration is to investigate
more data mining approaches alongside machine learning systems, and to express emerging collaboration
of web analytics with artificial intelligence.
Keywords Data Analytics, Sentiment Mining, Machine Learning
----------------------------------------************************---------------------------------I. INTRODUCTION
Throughout the years Web Analytics have
developed to quantify and help business requests
and research queries. Analytics assumes to be
essential part in the present time for the
advancement procedure of any association.
Presently a day's organizations are concentrating
more towards learning revelation & separating the
concealed data from vast information sets. While
doing the online shopping, swipe our card for
installment, shop something in bigger market places,
messaging/emailing, conferences and performing
different business assignments online over the web
we generate huge or we can say generating big data,
and the hidden part is that people are unaware of
that. Organizations use such kind of data to do the
market analysis and sets the trends as well as
change their policies such as changing cost,
customer satisfaction etc. usually they consider the
sentiment of number of visitors on site towards
particular block [1]. So such data helps the
companies to gather the statistics and do the market
analysis.

ISSN: 2394-2231

II. MULTISTEP APPROACH FOR WEB ANALYTICS PROCESS

The Web Analytics Process uses the multistep
approach in its first step collection of data is done,
and this task is done by performing association of
data items to get similar data of interest together.
This step provides the rules on which relations
between data items are established. Secondly,
preprocessing of data in information is done and the
data is transformed into information on the basis of
classification of data items. Variety of classification
techniques are existing like linear programming,
decision trees, neural networks etc. in the third step
the development of key performance indicators is
carried out, based on this heuristic information,
regression analysis done which sets the ground for
alternative business strategies. Finally formulation
of online strategy is done by evaluating
performance indicators. Web analytics assist the
formulation process by various models of machine
vision in web analytics. It can be expressed in
Figure 1.

http://www.ijctjournal.org

Page 205

International Journal of Computer Techniques - Volume 3 Issue 2, Mar-Apr 2016

IV.

Fig. 1 Web Data Analytics Process Model

III.

MACHINE LEARNING APPROACHES TO

WEB ANALYTICS
The main objective of machine learning is that, to
increase the accuracy of the algorithms and the
objectives of data mining are not different from
machine learning from practical point of view. Both
do not make efforts for collecting or managing or
pre-processing data, here the emphasis is just on
expanding the precision. Distinction is that, in data
mining extends above all else machine learning
calculations that are appropriate and versatile to
tremendous and monstrous datasets without
considering exactness are talked about. At the end
of the day in data mining ventures pertinence and
adaptability are more critical than exactness. The
various approaches are used to do such task these
are as follow:
Decision Tree Learning
Association Rules Learning
Artificial Neural Networks
Meta Imperative Learning
Support Vector Machines
Analysis through Boosted Trees
Clustering Process in Analysis

ISSN: 2394-2231

LITERATURE SURVEY

In 2001 G. Adomavicius et. al. [2] observed

personalization community by taking customers
profiles into account for finding out behavior of
customers and nature of their interactions. As a
model they have developed two parts of the whole
framework. One part contains data sets from
customer profiles and other contains rule base for
fitting these profiles under some logical processing.
In 2001 J. Su. et. al. [3] found medical data
discovery using three basic techniques named
Bayessian Network, Decision Trees and Back
Propagation Neural Networks. According to this
approach mined medical information is not only
classified for research perspectives but also made
available to the physician to improve his practices.
Six steps of knowledge discovery include data
collection firstly followed by data filtration, data
enhancement, data encoding, data mining and at last
knowledge representation. High correlation
parameter is chosen for construction of Bayessian
networks. While learning rate of 0.5 with learning
time of 1000 nanoseconds is taken for back
propagation technique. After generating decision
tree the results of three algorithms were discussed
by considering average of high accuracy of data
sets. In 2001 Y. Leung et. al.[1]
proposed
regression class mixture decomposition method for
mining regression classes in large data sets. They
have used iterative and genetic based optimized
objective function to guide regression curves. They
have found that it is not a real matter that only a
regression analysis can solve purpose of
identification in large data sets. For more accurately
modeling this identification process multiple
approaches have to be fit to explore classes from
large data sets. In November 2003, B. Fan et. al. [4]
proposed a spatial data mining method for web
analytics for evaluating customers intelligence.
The idea was to separate non spatial data of
customer from spatial data items. Non spatial data
of customer such as age of customer and the
population of a particular region whereas spatial
data comprises location of customer with locality as
a region. On the basis of data classification
techniques like k means clustering, spatial

http://www.ijctjournal.org

Page 206

International Journal of Computer Techniques - Volume 3 Issue 2, Mar-Apr 2016

overlapping method is designed for detecting

potential customers. In 2005 J. Xing et. al. [6] used
GIS as spatial data analysis tool which can
automatically discover implicit knowledge from
spatial data. Their idea was to develop web based
data mining system by integrating state of the art
GIS and data mining functionality in a closely
coupled open and extensible system architecture.
The idea was to extend predictive model markup
language (PMML) with spatial data mining to get
new version named SPMML. In 2006 W. Niblack
et. al. [8] allow uniform access to wide variety of
sources like bulletin boards, news feeds, reports by
analyst, a platform for very large scale text
applications that is known as web fountain. By
using set of hosted web services allows easy
development of documents in the form of text that
are useful for deriving end user applications. With
the help of application programming interfaces the
users are able to develop remote analytical
components, most useful application of web
fountain is reputation management. In 2006, S.
Yoon et. al. [9] [10] found usefulness of biochips
for acquisition of biological data with high
throughput. The main advantages of micro fluidic
lab on a chip include ease of use, speed of analysis,
low sample and reagent consumption, and high
reproducibility due to standardization and
automation. Without effective data-analysis
methods, however, the merit of acquiring massive
data through biochips will be marginal. In 2011 C.
Yang at. al. [7] found that analysis of developing
web opinions and social interaction is potentially
valuable for discovering ongoing topics of interests
of the public like terrorist and crime detection,
understanding how topics evolve together with the

underlying social interaction between participants,

and identifying important participants who have
great influence in various topics of discussions. In
2011, C. Wang et al. [5] observed that with the
expansion of web on the internet, it becomes the
great demand of times to analyse users sentiments
or opinions for the betterment of society and
humanity. In 2015 N. Koul et. al. [11] performs
survey of machine learning approaches to opinion
mining & sentiment analysis in text data. In 2016 I.
Cho et al [12] observed the task of identifying
relevant textual information on places, time periods,
people, and events is challenging since much
information on locations and times as well as their
relationships are described in many separate texts
with no clear linkage or structure. They have
proposed a visual analytics approach to help people
gain knowledge through making connections on
places, times, and events. The comparative analysis
of few literature surveys is expressed in Table1.
V. CONCLUSION
Data mining alongside with machine learning
systems for sentiment analysis is beginner approach.
This paper has observed that mining accuracy
through regression class mixture decomposition
over Gaussian mixture decomposition the results
would be more significant in customer profile
building, mining medical databases and intelligence
mining, if sentiment analysis of available data sets
is done. This paper also highlighted useful
application of data analytics for reputation
management in web fountain, and importance of
web opinion mining, limitations of existing
methods of web text mining. Focus was to highlight
various applications of web analytics.

TABLE I: Comparative Study Literature Survey

Methodology & Tool

Rule Based Classification

ANN & Fuzzy Logic ,

Decision tree Induction

ISSN: 2394-2231

Year
of
Publication
2001

Data Set Type/

Investigation Area
Building
Customer Profiles

Author

Results Discussion

G.
Adomavicius
et. al [2]

2001

Mining Regression
Classes

J. L. Su. Et.
al.[3]

In 2001 author observed personalization community

by taking customers profiles into account for
finding out behavior of customers and nature of their
interactions. As a model they have developed two
parts of the whole framework. One part contains
data sets from customer profiles and other contains
rule base for fitting these profiles under some logical
processing. These two parts are named by them as
data model and profile models.
In 2001 Author found medical data discovery using
three basic techniques named Bayessian Network,
Decision Trees and Back Propagation Neural

http://www.ijctjournal.org

Page 207

International Journal of Computer Techniques - Volume 3 Issue 2, Mar-Apr 2016

Regression analysis

2001

Mining
Medical
Databases

Y. Leung et.
al.[1]

REFERENCES
[1]

[2]

[3]

[4]

[5]

[6]
[7]

[8]

[9]

[10]

[11]

Yee Leung, Jiang-Hong Ma, & Wen-Xiu Zhang. (2001). A new

method for mining regression classes in large data sets. IEEE
Transactions on Pattern Analysis and Machine Intelligence, 23(1), 521. doi:10.1109/34.899943.
Adomavicius, G., & Tuzhilin, A. (2001). Using Data Mining Methods
to Build Customer Profiles. Data Mining and Knowledge Discovery,
5(1/2), 33-58. doi:10.1023/a:1009839827683.
Jenn-Lung Su, Guo-Zhen Wu, I-Pin Cha. (2001, October). THE
APPROACH OF DATA MINING METHODS FOR MEDICAL
DATABASE. Paper presented at IEEE, Istanbul, Turkey.
BO FAN, YI-JUN LI, XIANG-BIN YAN. (2003, November).
SPATIAL
DATA
MINING
METHOD
FOR CUSTOMER
INTELLIGENCE. Paper presented at IEEE, China.
Wang, C., Xiao, Z., Liu, Y., Xu, Y., Zhou, A., & Zhang, K. (2013).
SentiView: Sentiment Analysis and Visualization for Internet Popular
Topics. IEEE Transactions on Human-Machine Systems, 43(6), 620630. doi:10.1109/thms.2013.2285047.
Jeonghee Yi Wayne Niblack. (2005). Sentiment Mining in
WebFountain. San Jose, CA.
Yang, C. C., & Dorbin Ng, T. (2011). Analyzing and Visualizing Web
Opinion Development and Social Interactions With Density-Based
Clustering. IEEE Transactions on Systems, Man, and Cybernetics Part
A:
Systems
and
Humans,
41(6),
1144-1155.
doi:10.1109/tsmca.2011.2113334.
Jin Xingxing, Cai Yingkun, *Xie Kunqing, Ma Xiujun, Sun Yuxiang,
Cai Cuo. (2006, January). A Novel Method to Integrate Spatial Data
Mining and Geographic Information System. Paper presented at IEEE,
Beijing, China.
Wang, C., Xiao, Z., Liu, Y., Xu, Y., Zhou, A., & Zhang, K. (2013).
SentiView: Sentiment Analysis and Visualization for Internet Popular
Topics. IEEE Transactions on Human-Machine Systems, 43(6), 620630. doi:10.1109/thms.2013.2285047.
Yang, C. C., & Dorbin Ng, T. (2011). Analyzing and Visualizing Web
Opinion Development and Social Interactions With Density-Based
Clustering. IEEE Transactions on Systems, Man, and Cybernetics Part
A:
Systems
and
Humans,
41(6),
1144-1155.
doi:10.1109/tsmca.2011.2113334.
N. Koul, S. Hassan, A Survey of Machine Learning Approaches to
Opinion Mining & Sentiment Analysis in Text, International

ISSN: 2394-2231

[12]

[13]

[14]

[15]

[16]

Networks. According to this approach mined

medical information is not only classified for
research perspectives but also made available to the
physician to improve his practices. Six steps of
knowledge discovery include data collection firstly
followed by data filtration, data enhancement, data
encoding, data mining and at last knowledge
representation. High correlation parameter is chosen
for construction of Bayessian networks. While
learning rate of 0.5 with learning time of 1000
nanoseconds is taken for back propagation
technique. After generating decision tree the results
of three algorithms were discussed by considering
average of high accuracy of data sets.
Author proposed Regression Class Mixture
Decomposition (RCMD) method for mining
regression classes in large data sets. They have used
iterative and genetic based optimized objective
function to guide regression curves. They have
found that it is not a real matter that only a
regression analysis can solve purpose of
identification in large data sets. For more accurately
modeling this identification process multiple
approaches have to be fit to explore classes from
large data sets

Conference on Communication, Information and Computing

Technology, 12-13 May, 2015.
Cho, I., Dou, W., Wang, D. X., Sauda, E., & Ribarsky, W. (2016).
VAiRoma: A Visual Analytics System for Making Sense of Places,
Times, and Events in Roman History. IEEE Trans. Visual. Comput.
Graphics, 22(1), 210-219. doi:10.1109/tvcg.2015 .2467971
S. Sharma, A Review on Efficacy of Artificial Neural Networks in
Medical & Business Areas , International Journal of Recent Trends in
Engineering & Research, Volume 02, Issue 04, April 2016
S. Sharma, Cervical Cancer stage prediction using Decision Tree
approach of Machine Learning , International Journal of Advanced
Research in Computer and Communication Engineering, Vol. 5, Issue
4, April 2016.
S. Sharma, A Study on Data Mining Horizons, International Journal
of Recent Trends in Engineering & Research, Volume 02, Issue 04;
April 2016.
http://en.wikipedia.org/wiki/Data_mining

http://www.ijctjournal.org

Page 208