Exploring Data Mining Methodologies and Applications in Business

Desklib provides past papers and solved assignments for students. This report explores data mining techniques for business analytics.

Introduction
As in major sectors of industries, data science is getting more and more importance. So, at this
present challenging scenario, there is a lot of competition between different organizations. Every
business organization has mountains of data related to their strategies, sales, costs, customers,
networks and much more. As there is a deeper potential of business which is hidden behind this
large volume of data. Thus for business success, it is important to discover a credible pattern
related to this data. These patterns proved to be significant for success goals of the organization.
These patterns can be generated through data mining techniques employed by the authorities of
the organization for converting the raw data of the company into useful information so to
increase the sales, develop their more effective strategy, for increasing sales and company’s
growth. Hence, data mining becomes an integral part of the companies to gain their potential and
competitive edge in the global market.
DATA MINING
Data mining is also known as Discovery of knowledge in the database that is KDD which deals
with various techniques of analysis and handling the vast amount of data of the organization
(Shmueli, Bruce, Yahav and Patel, 2017). These techniques can be used to develop related novel
pattern between different streams of the organizational data for maintaining their better potential
and success in the business market. The user can analyze the data from different perceptions and
dimensions, and summarize the useful data through data mining (Kasemsap, 2015).
Data Warehouse
In daily operations of an organization, the conventional database systems have been used which
are also known as online transaction processing system (OLTP). These systems are responsible
for determining the business transactions in the data through queries and reporting tools provided
to the organization but these tools cannot provide any intelligence to the organization that why it
is happening. For this purpose, separate data warehouse has been generated which is also a
database that contains all the summary information collected from the OLTP systems in the
organization. Following the data mining techniques in this warehouse will help the authorities of

the organization in decision making by providing profitable information and segmenting their
customers across the world (Vaisman and Zimányi, 2014).
Data mining tasks
In data mining, there are several types for discovering different patterns interrelated to the
organizational data for specific goals and needs. Thus, data mining deals with a question that
what kind of pattern is useful for the organization. So, on the basis of kind of data, there can be
two kinds of functions in data mining:
 Descriptive mining: This task involved in generating a description and general
properties of the data stored in the warehouse database. These functions involve the
following:
 Concept and mining: In this type of mining, the categories have been identified
that which category holds which type of data. The class of the data and the
concept behind it can be mined through this technique.
 Anomaly detection: As a large data set, there can be possibilities of the data to be
an anomaly with its respective pattern. This means the pattern whose behavior is
not normal in the dataset has been identified in this context. So, by statistical
processes, the organization can determine something which is notably different in
the environment (Agarwal and Agarwal, 2015)
 Classification: In this technique, different data forming different pattern have been
classified and categorized. Different classes and categories have been identified to
retrieve only the relevant and related data.
 Association learning: In this type of data mining, the data about the associated
recommended systems have been generated and analyzed. Taking an example of
customers buying cocktail shaker and recipe book also often buy the glasses along with
it. This type of data analysis for associating different systems and co-relating them for
business profit statement comes under associative data mining.
 Cluster detection: This process helps in determining one type of pattern through the
given data set by recognizing the clusters or subcategories of the data in that large data
set. As purchasing habits of different people vary person to person, so by analyzing the

purchasing behavior of different groups and subcategorizing it within different data set
will help the organization for making business strategies. The differences and similarities
between the data can be recognized by this technique (Nadiammai and Hemalatha, 2014).
 Regression: By identifying and analyzing the relationship between different variables is
done in this technique. The presence of one variable is appreciated in this technique
regardless of the existence of another variable.
 Prediction: This type of data mining involves making predictions among the current
gathered data and with help of some variables and fields related to the data set, future
values of interest and variables have been predicted. A model is being developed for
achieving the future outcomes on the basis of passed records of data sets with known
fields. In this process, the patterns detected and structured by the descriptive and
classified data mining can be used for further findings to achieve the goals of the
organization (Mishra, Kumar and Gupta, 2014).
Methodologies used in data mining
 Decision tree (DT): This methodology is widely used in decision analysis where the node
or terminal represents the decision that is right or wrong and the segments denote the
further processing. The criteria of choosing any one branch depend upon the outcome of
the test. Based on the data, predictions have been generated (Rutkowski, Jaworski and
Duda, 2014). The business organizations highly rely on decision trees for solving the
problems and answering the questions. Taking an example from our day to day life that is
the weather good enough so to go outside? There can be yes and no answers of this
simple question but also, at the same time, there can be various possibilities of sub-
questions and categorization of data. Hence, this decision tree is so called the best
methodology for mining of the large data sets but firstly a summarized form will be input
to this methodology (Ahmed and Elaraby, 2014).
 Neural Networks: To find the patterns in the data and the complex relationship between
input and output data, then this non-linear and statistical data modeling tool comes into
the scenario. The data processing including filtering, clustering, blind source separation,
and compression can tend to be done with neural networks. As the neural network is

characterized as good robustness functionality, distributed storage, adaptive and self-
organized function. As the neural network programmed to find patterns, store, recognize
and retrieve the pattern and data entries. Hence in many ways, the neural network
methodology serves the data mining in the organization.
 Genetic algorithm: By combining some good parts of all the expected solutions, a better
solution could be generated. This approach is known as a genetic algorithm. This
technique in the case of data mining can be utilized for selection and gives optimized
solutions (Ali and Tuteja, 2014).
Data mining implementation process
The main phases in data mining are provided below. There can be moves for back and forth in
between the different phases which are:
 Business understanding: By understanding the project objective and the requirement from
the perspective of the business that which information is needed throughout the database
which is important for the growth of the company. By understanding the business
objectives as well as the client’s objective and then by analyzing current data scenario,
data mining goals can be achieved.
 Data understanding: As the data can be collected from different sources available.
Starting from data collection and then proceeding further to identify the quality problems
of data, or to develop the interesting subsets, comes under this phase. By using the query,
decision tree, and visualization tools, data mining questions can be answered.
 Preparation of data: All the data collected in the previous case can be constructed to a
final dataset in this phase. In this task tables, selection of attributes, records, and cleaning
of data modeling tools have been included so that a proper role model will be decided to
be operated upon. The data from the different phases can be stored, selected, cleaned,
formatted, transformed and constructed. This phase is very much important in the data
mining project (Furnas, 2012).
 Modeling: On the basis of business objectives, suitable modeling techniques should be
imported on the prepared dataset in the previous phase. The stakeholders of the business
organization should make sure that the model meets the data mining objectives.

 Evaluation: In this phase, the results have been evaluated which are generated through the
models of data mining and can be fetched against the business strategies.
 Deployment: In this phase, the data mining discoveries have been implemented in the
business environment for attracting the consumers more and more for a profitable
statement of the organization. A final report has been created in this phase for further
improving the business policies. A detailed plan has also been deployed for monitoring
the data mining discoveries that have been created (Furnas, 2012).
Conclusion
In this report, we have discussed the data mining techniques. At present time, a wide community
of data analyst has to deal with the discussed data mining techniques. For identifying different
patterns between the data sets of the organization and improving the vulnerability can lead the
business to grow smoothly. We have also discussed the different methodologies which help the
authorities to perform data mining in which neural network is the main and import method. By
using this method in data mining, the complex relationship can also be analyzed between the
input and output assets of the organization. The lifecycle of data mining have also been discussed
here in which the business and information understanding is done prior in the organization and
after that development or manipulation in data, development of different models and the
evaluation of these models have been done for proper deployment of data mining in the
organization.

References
Kasemsap, K., 2015. The role of data mining for business intelligence in knowledge
management. In Integration of data mining in business intelligence systems (pp. 12-33). IGI
Global.
Vaisman, A. and Zimányi, E., 2014. Data warehouse systems. Data-Centric Systems and
Applications.
Ali, S.M. and Tuteja, M.R., 2014. Data Mining Techniques.
Ahmed, A.B.E.D. and Elaraby, I.S., 2014. Data Mining: A prediction for Student's Performance
Using Classification Method. World Journal of Computer Application and Technology, 2(2),
pp.43-47.
Nadiammai, G.V. and Hemalatha, M., 2014. Effective approach toward Intrusion Detection
System using data mining techniques. Egyptian Informatics Journal, 15(1), pp.37-50.
Agrawal, S. and Agrawal, J., 2015. Survey on anomaly detection using data mining
techniques. Procedia Computer Science, 60, pp.708-713.
Furnas, A. (2012). Everything You Wanted to Know About Data Mining but Were Afraid to Ask.
[online] The Atlantic. Available at:
https://www.theatlantic.com/technology/archive/2012/04/everything-you-wanted-to-know-about-
data-mining-but-were-afraid-to-ask/255388/ [Accessed 25 Jan. 2019].
Shmueli, G., Bruce, P.C., Yahav, I., Patel, N.R. and Lichtendahl Jr, K.C., 2017. Data mining for
business analytics: concepts, techniques, and applications in R. John Wiley & Sons.
Rutkowski, L., Jaworski, M., Pietruczuk, L. and Duda, P., 2014. Decision trees for mining data
streams based on the gaussian approximation. IEEE Transactions on Knowledge and Data
Engineering, 26(1), pp.108-119.
Buczak, A.L. and Guven, E., 2016. A survey of data mining and machine learning methods for
cyber security intrusion detection. IEEE Communications Surveys & Tutorials, 18(2), pp.1153-
1176.

Mishra, T., Kumar, D. and Gupta, S., 2014, February. Mining students' data for prediction
performance. In Advanced Computing & Communication Technologies (ACCT), 2014 Fourth
International Conference on (pp. 255-262). IEEE.