See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/270338372

Overview on Data Mining Schemes to Design Business Intelligence

Framework for Mobile Technology

Article · December 2014

CITATION READS

1 962

3 authors, including:

Manish Tembhurkar Dr-Ashlesha S Nagdive

Raisoni Group of Institutions Raisoni Group of Institutions
14 PUBLICATIONS 33 CITATIONS 13 PUBLICATIONS 75 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Manish Tembhurkar on 03 January 2015.

The user has requested enhancement of the downloaded file.

 ISSN No. 0976-5697
Volume 5, No. 8, Nov-Dec 2014
International Journal of Advanced Research in Computer Science
REVIEW ARTICLE
Available Online at www.ijarcs.info

Overview on Data Mining Schemes to Design Business Intelligence Framework for

Mobile Technology
Manish P. Tembhurkar Dr. R. M. Tugnayat
Department of CSE, Principal,
G. H. Raisoni College of Engineering, S. S. Agnihotri College of Engineering,
Nagpur, India Wardha, India

Ashlesha S. Nagdive
Department of IT,
G. H. Raisoni College of Engineering,
Nagpur, India

Abstract: Mining is such as automated technology which facilitates us to find the relations between various parameters and predicts the future
outcomes of business sector. Mining tools (data mining, text mining, and web mining) are used to foreseen the future, profit, investment relations
in large databases for BI. In this paper, we review various data mining tools and techniques and their performance in BI sectors. Finally, we shall
state some of the important factors to be considered while designing & developing a BI framework using data mining techniques for various
sectors.
Business Intelligence (BI) has turned out to be a predictable technological advantage in the last couple of years, for the large enterprises
(especially Mobile Technology Ventures). Such ventures could afford to buy, implement and maintain BI solutions. Recently, small and medium
size enterprises all over the globe have understood the competitive and financial benefits of BI. However, limited IT budgets of small companies
and BI’s high total cost of ownership have created a gap between large and small enterprises where small enterprises do not become fortunate to
avail the virtues of BI because of the affordability factor.
This Paper provides the systematic study and analysis of various factors (Data Mining techniques, BI framework and their effects on designing a
successful automated decision making system). It presents the strong knowledge foundation so that one can give a proper justice to achieve BI
objectives.

Keywords: Data Mining, Data visualization, Data warehousing, Mobile Technology, Nanocube, Spatiotemporal dataset, OLAP, Business
Intelligence.

I. INTRODUCTION analyzing and leveraging of so that it can be converted into

to easily understandable structure. The objective of BI is to
A. Data Mining: assist a user to make easy decision to assist their business.
Data mining is the technology comprises of tools and C. BI Terms [3]:
techniques for the extraction and analysis of knowledge in
systematic format from the large database. Sometimes, the There are several BI terms currently in discussion such
large amount of data is very hard, complex and time as,
consuming to process and derive the expected outcome from operational databases,
it. Larger the size if database, larger will be the processing OLTP,
overload, delay for processing. Furthermore, the extraction Data warehouse,
of exact required knowledge for analyse is crucial when it data mart,
comes for the BI sector. The prior objective of any BI is to OLAP and OLAP Server,
achieve business patterns. These BI patterns [1] are further Drill down/up,
being processed and analysed for designing new business ad-hoc query and analysis,
goals. Data mining uses tools and techniques for the dimension,
following [2]. Fact/Measure,
Prediction KPI,
Classification dashboard, and
Segmentation/clustering scorecard.
Affinity analysis using association rules.
D. BI Tools [3]:
B. Business Intelligence [3]: BI tools are the tools & techniques in the form of
BI can be defined as the collection of tools and application software designed to
techniques. BI transforms raw data into significant Report,
information useful for analyzing business systems. Business Analyze and
Intelligence refers to the process of collection, structuring, Present data.

© 2010-14, IJARCS All Rights Reserved 128

 Manish P. Tembhurkar et al, International Journal of Advanced Research in Computer Science, 5 (8), Nov–Dec, 2014,128-133

The BI techniques primarily read data that have been B. Classification:

provided, though a data warehouse or data mart. BI tools can
be categorised into the following three categories: Data Mining Models
a) Query and reporting;
b) Online analytical processing (OLAP); and
c) Information mining/ Data Mining Predictive

II. DATAMINING ALGORITHMS

Association Rule
Data recovery operations and future prediction
operations are based on the data mining techniques used.
Figure 2 demonstrates some of the significant techniques Apriori Algorithm
and algorithms to produce and learn fascinating system
patterns.
The Data warehousing [4] & Data mining techniques Eclat Algorithm
[5] are initially classified as either Predictive or Descriptive
Data Mining Models. In addition, we further provide
comparative Analysis of four Data Mining Techniques & FP- growth Algorithm
Algorithms in Table I. These four techniques are;
a) Nanocubes
b) Association Rule Mining Clustering
c) Classification
d) Regression
Definitions and limitations are stated for these K-Means
respective Data Mining techniques in Table I.
A. Data Mining with Spatiotemporal Datasets
Bi-Clustering
(Nanocubes) [11]:
Nanocubes provide a possible solution for an efficient
storage and to access high dimensional datasets shown in Density Based
figure. As datasets get larger, exploratory data visualization
becomes more difficult. Consider a dataset with a billion
entries. We can compute a small summary of the dataset and Agglomerative Hierarchic
visualize the summary instead of the dataset, summaries
might help, but in order to understand if that is the case, we
will inevitably find ourselves having to visualize over million Descriptive
residuals.

Classification

C 4.5

K-Nearest Neighbor Algorithm

Naïve Bayes

ADA BOOS

Regression

Generalized Linear Models

Support Vector Machine

Figure 2. Classification of Data Mining Algorithms

Figure 1. Data Visualization with Data Cubes (3 dimensions: Type,
Location & Time) [16]

© 2010-14, IJARCS All Rights Reserved 129

 Manish P. Tembhurkar et al, International Journal of Advanced Research in Computer Science, 5 (8), Nov–Dec, 2014,128-133

Table I. Comparative Analysis of Data Mining Techniques & Algorithms

Sr. No. Techniques Algorithm Name Introduction Limitations
• Timings for the queries are
• Data cubes are structures that perform dominated by network and user-
aggregations across every possible set of interaction latencies. (WM)
OLAP Datacubes
1 NANOCUBES dimensions of a table in a database, to support • specifically to answer queries from
[8]
quick exploration. [7] interactive visualization systems
• High Dimensional Data Visualization [16] that explore massive datasets [7]

• Apriori Algorithm find frequent item sets from a • Candidate generation generates
ASSOCIATION Apriori
2 transaction dataset and derive association rules large numbers of subsets. [12]
RULE MINING Algorithm [12]
[12]
• C4.5 is an algorithm used to generate a decision
• Does not work well with small
tree. [13]
training data set [13]
• Such systems take as input a collection of cases,
• Small variation in data can lead to
each belonging to one of a small number of
3 CLASSIFICATION C4.5 different decision trees (especially
classes and described by its values for a fixed set
when the variables are close to each
of attributes, and output a classifier that can
other in value)
accurately predict the class to which a new case
belongs.
• In machine learning, support vector are supervised • One of the initial drawbacks of
learning models with associated learning SVM is its computational
algorithms that analyze data and recognize inefficiency [14]
patterns, used for classification and regression • SVM is a binary classifier. To do a
Support Vector analysis. [14]
4 REGRESSION multi-class classification, pair-wise
Machines
• The aim of SVM is to find the best classification classifications
function • can be used (one class against all
• to distinguish between members of the two classes others, for all classes
in the training data.

III. REQUIREMENT OF BI IN MOBILE f. Image sensor

TECHNOLOGY g. Display
h. Mobile operating system*
Data mining provides overviews, filters, zooming, and A. Mobile Innovations Forecast[6]:
details-on-demand inside the spatiotemporal dimensions
Questions in Mobile Technology require not just a keen
themselves. But three stages are to be achieved for using the
understanding of the innovative approaches of the enabling
technique of Data Cubes [8] before we proceed on further
objectives: technologies, but also a wider approach for analysing mobile
a. Design OLAP Datacubes innovation quantitatively and qualitatively. Four-part
frameworks for analysing and understanding mobile
b. Data Aggregation
innovation are:
c. Data Visualization
a. Enabling technologies;
b. New technological capabilities;
Whereas the Mobile Technologies Index [6] is
concerned, a broad composite of eight enabling components c. New use cases and
that underlie the power of the mobile device & d. New business models.
Quantitative methods are the techniques developed by
communication [9] to sense, analyse, store and connect
the BI sector to analyse the exact rate of improvement as a
information for Business Intelligence sector. The eight
key factor. Such Technologies are fundamentally crucial to
components are as follows:
mobile innovations, and to assist predicting new use cases
a. Device connectivity speed
b. Infrastructure speed and business models.
c. Processor speed
d. Memory
e. Storage

© 2010-14, IJARCS All Rights Reserved 130

 Manish P. Tembhurkar et al, International Journal of Advanced Research in Computer Science, 5 (8), Nov–Dec, 2014,128-133

.
Figure 3. Mobile technologies index - relative progress of components (source: IHS Supply Database) [15]

Table II. Top 30 Mobile Service Provider companies in the world

c) There is also a growing need from industry players
(Source: RBI, Financial Access 2013-CGAP) [15] such as banks, educational institutions, healthcare
providers etc. to use mobile applications and other
Total
Sr.
Company Country subscribers
value added services as a channel to provide easy
No. access to services to their customers, as well as,
(Millions)
1. China Mobile China 775.6 increase productivity and efficiency of employees.
2. Vodafone UK 419.4 d) MVAS is also being seen as a means for reaching
3. China Unicom China 285.7 customers who were previously inaccessible due to
4. Airtel India 275.2
5. América Móvil Mexico 269.9 lack of the presence of physical infrastructure and/
6. Telefónica Spain 254.7 or high cost of servicing in remote areas.
7. Axiata Malaysia 239.7
8. Orange France 231.5
9.
10.
VimpelCom Ltd.
China Telecom
Russia
China
209
185 Total
11. MTN Group South Africa 175.98 3000
12. Etisalat UAE 167 2500
13. Telenor Norway 166
Swedan/ 2000
14. TeliaSonera 160
Finland 1500
15. T-Mobile Germany 142.5
16. Saudi Telecom Company (STC) Saudi Arabia 139 1000 Total
17. Reliance Communications India 135.88
500
18. Verizon Wireless USA 122
19. Idea Cellular India 113.9 0
20. AT&T Mobility USA 116.01
21. MTS Russia 106.07 Africa Asia Europe North
22. Telecom Italia Mobile (TIM) Italy 102.5 America
23. BSNL India 96.28
24. Tata Teleservices India 77.4
25. Turkcell Turkey 70.7 Figure 4. Number of subscribers (in millions) of region
26. Aircel India 66.79 We believe there is a vast scope for the research in Mobile
27. Smart Communications Philippines 72.5
28. Maxis Communications Malaysia 63.71
Technology here in India. There are various positive aspects
29. MegaFon Russia 62 about the country, such as,
30. Ooredoo Qatar 60.53 a. India has the 2nd largest population in the world,
hence Mobile users cover 2nd largest portion of the
a) Automation due to Information and total population of the world.
Communications Technology (ICT) i. Total number of mobile phones in India:
b) Business needs of service providers 904,510,000
ii. Total Population of the country: 1,220,800,359

© 2010-14, IJARCS All Rights Reserved 131

 Manish P. Tembhurkar et al, International Journal of Advanced Research in Computer Science, 5 (8), Nov–Dec, 2014,128-133

iii. Ranks with respect to mobile phone users in the UAE 167
world: 2nd Europe 1924.37
iv. phones per 100 citizens: 74.09 mobile phones per
100 citizens France 231.5
b. 13 mobile service providers in the country Germany 142.5
c. 6 Indian Mobile providers Operators have secured Italy 102.5
their positions in "top 30 Mobile Network Operators
Norway 166
in the world" list
d. 2nd Highest no. of mobile phone users in the world Russia 377.07
after China. Spain 254.7
i. Total no. of mobile users in India: 765.45
Swedan/ Finland 160
millions
ii. Total no. of mobile users in the world: 5362.45 Turkey 70.7
millions UK 419.4
iii. Ratio: 1/7 North America 507.91
Table III. Number of Subscribers (in millions) in region (Countrywise) Mexico 269.9
Region/ Country Sum of Total subscribers (Millions) USA 238.01
Africa 175.98 Grand Total 5362.45
South Africa 175.98
IV. DATA TRANSFORMING INTO BI
Asia 2754.19
China 1246.3 Data Transforming into BI is a revolutionary approach
India 765.45 to understand the potentials of business enterprises and to test
their limits. This approach provides the completive
Malaysia 239.7
advantages; understand the hidden relations between various
Malaysia 63.71 business components. Every system has to have a systematic
Philippines 72.5 approach to achieve its ultimate goal. Implementing Data
Qatar 60.53
mining in BI is no other exception.
Saudi Arabia 139

Figure 5. Data Transforming into BI [2]

Following are the five important stages (as shown in C. Data warehousing
Figure 5) to apply data transforming into BI successfully. D. Implementation of BI tools
A. Collection of Raw data from Business Enterprise E. Analyzing outputs
B. Data Cleaning through search Engines & Filtering Some of the BI tools in the market are,
process • SAP Business Objects [17];

© 2010-14, IJARCS All Rights Reserved 132

 Manish P. Tembhurkar et al, International Journal of Advanced Research in Computer Science, 5 (8), Nov–Dec, 2014,128-133

• IBM Cognos [18]; [7] LauroLins, James T. Klosowski, and Carlos Scheidegger,
• ProClarity [19]; “Nanocubes for Real-Time Exploration of Spatiotemporal
• and QlikView [20] Datasets”, IEEE Transactions on Visualization And
Computer Graphics, Vol. 19, No. 12, December 2013.
V. CONCLUSION
[8] Abello, A. ; Romero, O. ; Pedersen, T. ; Berlanga Llavori,
Business sector is very dynamic in nature and BI tools R. ; Nebot, V. ; Aramburu, M. ; Simitsis, A., “Using
provide us the BI solutions to keep up with rapidly evolving Semantic Web Technologies for Exploratory OLAP: A
sector. When the Data Mining technique was first used for Survey”, IEEE Transactions on Knowledge and Data
finding BI solutions, the approach opens the door for vast Engineering, Volume: PP, Issue: 99, 2014.
possibilities in future business sector. The Mobile [9] Deepak Pareek, “Book: Business Intelligence for
Technology sector is currently a booming sector. Data Telecommunications”, CRC Press, 29-Nov-2006, ISBN 0-
mining assists us to calculate E-Business solutions for this 8493-8792-2. Retrieved 18 March 2008.
growing sector. In this paper, we provide a systematic
[10] N. Elmqvist and J. D. Fekete, “Hierarchical aggregation
overview of three important domains (Data Mining, Business
Intelligence & Mobile Technology). The paper will surely for information visualization: Overview, techniques, and
assist to help designing the BI framework to understand the design guidelines”, IEEE Transactions on Visualization
Mobile Technology components. To find BI relations and Computer Graphics, 16(3):439–454, May 2010.
requires BI framework. Data mining techniques are the key [11] “Real – Time exploration of spatiotemporal datasets”
component of such framework. Official web site
http://www.nanocubes.net
VI. ACKNOWLEDGMENT
[12] Rakesh Agrawal and Ramakrishnan Srikant “Fast
One of us (M. P. Tembhurkar.) thanks RBI, Financial algorithms for mining association rules in large
Access 2013-CGAP for providing Mobile consumer database databases”. Proceedings of the 20th International
& Raman Chitkara, Founder of Mobile Innovations Forecast Conference on Very Large Data Bases, VLDB, pages 487-
for guiding us through the necessity of BI in Mobile 499, Santiago, Chile, September 1994.
Technology sector. [13] Wei Dai, Wei Ji, “A MapReduce Implementation of C4.5
Decision Tree Algorithm”, International Journal of
VII. REFERENCES
Database Theory and Application, vol 7, issue 1 pp. 49-
[1] Mohamed Ridda Laouar and Sean B. Eom, “Book: 60, 2014
Business Intelligence and Mobile Technology Research: [14] Max Welling, “Support Vector Machines”, Department of
An Information Systems Engineering Perspective”, Computer Science, University of toranto,
Cambridge Scholars Publishing, ISBN (10): 1-4438-5507- http://www.ics.uci.edu/~welling/classnotes/papers_class/S
3, ISBN (13): 978-1-4438-5507-5, 2014. VM.pdf
[2] Abdul-Aziz Rashid Al-Azmi, “Data, Text, And Web [15] “IHS iSuppli Mobile and Wireless Communication
Mining For Business Intelligence: A Survey”, Service” official Website
International Journal of Data Mining & Knowledge
https://technology.ihs.com/
Management Process (IJDKP) Vol.3, No.2, March 2013
[16] Vinod S. Bawane, Shireesh P. Bhoyar, Manish
[3] Muhammad Sohail, Khalid Khan, Rizwan Iqbal and
Tembhurkar, “A Review on High Dimensional Data
Noman Hasany, “Developing Business Intelligence
Visualization”, International Journal of Emerging Trends
Framework to Automate Data Mapping, Validation, and
in Engineering and Development (IJETED 2014), Issue 4,
Data Loading from User Application”, International
Vol.3, Page 878-884, ISSN 2249-6149, R S. Publication
Journal of Computer Theory and Engineering, Vol. 4, No.
(rspublication.com), May 2014.
5, October 2012.
[17] “SAP Business Objects” official website
[4] Risi, M. ; Sessa, M.I. ; Tucci, M. ; Tortora, G., “CoDe
Modeling of Graph Composition for Data Warehouse https://www.sapbusinessobjectsbi.com/
Report Visualization”, IEEE Transactions on Knowledge [18] “IBM Cognos software” official website
and Data Engineering, Volume:26 , Issue: 3, 2014. http://www.ibm.com/software/analytics/cognos/
[5] Zeljko Panian, “The Evolution of Business Intelligence: [19] “ProClarity - XLCubed” official website by Microsoft
From Historical Data Mining to Mobile and Location-
http://www.xlcubed.com/proclarity/
based Intelligence”, WSEAS International Conference on
Recent Researches in Business and Economics, ISBN: [20] “Qlik: Business Intelligence and Data Visualization
978-1-61804-102-9, 2012 Software” official website

[6] Raman Chitkara, “Mobile Innovations Forecast: The http://www.qlik.com/

elements of contextual intelligence”, Mobile Innovations
Forecast: Phase II, PwC, 2014

© 2010-14, IJARCS All Rights Reserved 133

View publication stats