Data Mining Methods for Business

This document discusses the common types of data mining methods used in business, such as association, classification, clustering analysis, prediction, pattern tracking, decision trees, neural network, and outlier analysis. It explains how these methods help organizations extract useful information from large data sets and make effective decisions. Examples are provided for each method.

DATA HANDLING
1

CONTENTS
INTRODUCTION...........................................................................................................................3
PART 1............................................................................................................................................3
Determine the sales and profit over years and evaluate the use of Excel for pre-processing the
data...............................................................................................................................................3
Demonstrate that how it can practically with use of Excel function such as if, Lookup, Pivot
table, charts and graphs................................................................................................................6
PART 2............................................................................................................................................9
Common types of data mining methods that can be used within a business...............................9
Advantages and disadvantages of Weka over Excel.................................................................12
CONCLUSION..............................................................................................................................13
REFERENCES..............................................................................................................................14
2

INTRODUCTION
Data mining is based on the process of finding the patterns, correlations and anomalies
within large data set. In order to predict accurate result or outcome. The data mining process will
support for marketing assistant to increase the revenue of information, cut cost while improving
the customer relationship. Afterwards, it will reduce the various risks and so on. Data mining is
consider as important factor which consider the large volume of data or information. In another
word, data mining is exploring and analyzing the large amount of data which easily discover
meaningful pattern, rules and facts. It will consider as a discipline under specific data science
and offer from predictive analytics. Furthermore, it will described the sales information while
primary aim is to predict future result or outcome. In additional, data mining is an appropriate
technique to build machine learning model that power in term of artificial intelligence.
PART 1
Determine the sales and profit over years and evaluate the use of Excel for pre-processing the
data.
It is based on the data mining technique always support for calculating the sales and
profit over years. It provides a brief idea about the critical success of modern data drive
organization. It is powerful tool that transform enterprise, however, it is also implementing the
process to meet the specific need of business. Sometimes, it is frequently stand in the better way
to successful process data mining.
Excel is consider as standard for spreadsheet software that enable to calculate the budget,
financial forecasting and various data entries. The primary function is to organise large amount
of data or information in proper manner. Furthermore, it can easily analyse the data to create
graphs, charts. In order to represent in visualise manner.
Row Labels Average of Profit Sum of Sales
Furniture 68.11660673 5178590.542
2009 140.1369955 1469508.194
2010 20.65391403 1250043.046
2011 115.326226 1258336.514
2012 -5.173357143 1200702.788
Office Supplies 112.3690738 3752762.1
3

2009 153.4285381 1031244.56
2010 97.14263473 885095.79
2011 80.42802855 816902.13
2012 117.6447423 1019519.62
Technology 429.2075157 5984248.182
2009 337.0125974 1668572.052
2010 474.5130402 1416503.546
2011 518.2162105 1380213.417
2012 398.3725568 1518959.168
Grand Total 181.1844243 14915600.82
Figure 1
Data Interpretation
By using excel, it can be calculating the Average profit and sum of sales. It can be
identified the profit and sales of furniture. Furthermore, it also identified the average profit in
year wise so that it become easier to identify the accurate information of furniture.
4

Figure 2
Data Interpretation
By using excel, it can be calculating the Average profit and sum of sales. It can be
identified the profit and sales of Office supplier. Furthermore, it also identified the average
profit in year wise so that it become easier to identify the accurate information of supplier.
Figure 3
5

Data Interpretation
By using excel, it can be calculating the Average profit and sum of sales. It can be
identified the profit and sales of technology. Furthermore, it also identified the average profit in
year wise so that it become easier to identify the accurate information of technology.
Demonstrate that how it can practically with use of Excel function such as if, Lookup, Pivot
table, charts and graphs.
Look up: The Look up function is one of lookup and reference function. It is mainly used
to return the value from selected range. It is consider as value to search the particular data into
large data set. Lookup vector is single row or column which containing sorted data in ascending
order. At the most of level. Look up function in excel search a particular value in one column,
rows which return a matching value from same position in another column. In another context,
vector refer to one column and one row range. As consequently, it can be used the vector from of
Lookup to search particular row or column for specified valued.
Syntax:
Lookup (lookup_value, lookup_vector, [result_vector])
6

Pivot table: it can be summarised the data by using Excel tool and extract data in proper manner.
Pivot table is consider as most suitable for summarise, recognise, count and calculate average of
data stored in the database. It allows for user to transform column into rows in proper manner.
Calculate the Sum of shipping cost, sum or product base margin and sum of sales for Furniture.
Row Labels
Sum of Shipping
Cost
Sum of Product Base
Margin
Sum of
Sales
Furniture 53243.69 1006.77 5178590.542
Bookcases 8646.07 122.09 822652.04
Chairs & Chairmats 15512.69 228.46 1761836.55
Office Furnishings 8402.72 414.31 698093.81
Tables 20682.21 241.91 1896008.142
Figure 4
Calculate the Sum of shipping cost, sum or product base margin and sum of sales for office
supplies.
Office Supplies 36095.51
Appliances 6854.11
Binders and Binder Accessories 6633.52
Envelopes 1682.77
Labels 288.66
Paper 7914.41
Pens & Art Supplies 2041.81
Rubber Bands 225.21
7

Scissors, Rulers and Trimmers 670.51
Storage & Organization 9784.51
Figure 5
Calculate the Sum of shipping cost, sum or product base margin and sum of sales for office
supplies.
Technology 18491.84 1148.77 5984248.182
Computer Peripherals 4067.34 449.67 795875.94
Copiers and Fax [phone] 1130361.3
Office Machines 7135.91 149.75 2168697.14
Telephones and Communication 4841.71 512.98 1889313.802
Figure 6
8

Calculate Sum of region By Profit
Row Labels
Sum of
Profit
Atlantic 238960.66
North Carolina 2841.11
Northwest
Territories 8307.05
Ontario 439214.57
Prarie 321160.12
Quebec 140426.65
West 297008.61
Yukon 73849.21
Grand Total 1521767.98
Figure 7
Calculate region wise order date, customer name and sum of order id
Region Order Date Customer Name
Sum of Order
ID
Atlantic 13/01/2011 Cynthia Voltz 55815
Atlantic 13/02/2009 Sally Matthias 33224
Atlantic 13/02/2010 Maureen Gastineau 59619
Atlantic 13/02/2011 Astrea Jones 10245
Atlantic 13/02/2011 Chad Cunningham 59045
Atlantic 13/02/2011 Jennifer Braxton 19074
Atlantic 13/03/2010 Carlos Meador 7938
Atlantic 13/04/2012 Jim Epp 102404
Atlantic 13/05/2009 Carlos Meador 13941
9

Atlantic 13/05/2011 Erica Bern 80654
Atlantic 13/05/2012 Benjamin Venier 15399
Atlantic 13/05/2012 Fred McMath 11011
Atlantic 13/05/2012 Michael Kennedy 35396
Atlantic 13/06/2009 Maribeth Dona 37348
Atlantic 13/06/2011 Helen Abelman 30048
Atlantic 13/07/2012 Karl Brown 14215
Atlantic 13/08/2010 Sean Wendt 5061
Atlantic 13/09/2011 Joy Smith 81742
Atlantic 13/10/2010 Dave Poirier 26401
Atlantic 13/12/2010 Sarah Bern 46849
Atlantic 14/01/2010 Dianna Wilson 58599
Atlantic 14/01/2011 Charles Sheldon 21379
Atlantic 14/02/2010 Muhammed Yedwab 24518
Atlantic 14/02/2011 Bradley Talbott 50276
Atlantic 14/02/2012 Alex Grayson 54053
Atlantic 14/02/2012 Thea Hendricks 1154
Atlantic 14/04/2009 Becky Martin 8422
Atlantic 14/04/2009 David Flashing 32999
Atlantic 14/04/2011 Roy French 9216
Atlantic 14/04/2012 Jack Lebron 38310
Atlantic 14/05/2009 Brad Thomas 36609
Atlantic 14/05/2010 Bill Tyler 22338
Atlantic 14/05/2010 Karen Seio 21004
Atlantic 14/06/2011 Frank Atkinson 44292
Atlantic 14/07/2011 Barry Franz 20676
Atlantic 14/07/2012 Lori Olson 14980
Atlantic 14/08/2009 Elizabeth Moffitt 15971
Atlantic 14/08/2011 Ionia McGrath 47810
Atlantic 14/08/2012 Katherine Hughes 102350
Atlantic 14/09/2009 Debra Catini 7297
Atlantic 14/09/2009 Muhammed Yedwab 7269
Atlantic 14/09/2012 Maureen Gastineau 84612
10

PART 2
Audileadership data in conjunction with Weka with the help of clustering
Figure 8 Current relationship of Audileadership
Figure 9 farthest first Clustering
11

Figure 10 kmeans Algorithm for Clustering
12

Common types of data mining methods that can be used within a business
Data mining helps in bringing and identifying useful patterns so that desired conclusion
can be drawn from data (Kavakiotis and et. al., 2017). There are various kinds of data mining
methods that are used by organizations for various kinds of purposes. These methods help the
organizations to extract required and useful information from large amount of data so that an
effective decision can be taken. Each of these methods have different purposes and are used as
13

per the requirement of the business. These methods also help the organizations in analyse current
market trend so that effective decisions can be taken in order to enhance overall revenue of the
organization. Some of the most commonly used data mining methods that can be used by
business are: association, Classification, clustering analysis, prediction, pattern tracking or
sequential pattern, decision trees, neural network and Outlier Analysis or Anomaly Analysis.
Each of them will be explained with the help of an example.
Association: It is one of the most common method which is used to find correlation between two
or more than two items so that a hidden pattern within them can be identified and because of this
it is also known as relation analysis. This method is mostly used by organizations so that they
can identify or predict behaviour of the customers or consumers (Bandaru, Ng and Deb, 2017).
This method is mostly used in supermarket so that they can identify which product is purchased
most frequently by the customers. For example: analysis of consumer data shows that if a
consumer buys milk then there are 70 percent chances that they will be buying bread with it and
if overall tractions are compared then it can be identified that more than 10 percent people have
purchased milk and bread together.
Classification: It is a kind of data mining method which is used to identify or distinguish
between items within a data set so that they can be classified or grouped within a category. It
also helps in predicting behaviour of the item within a group. It is completed in two steps: first is
the leaning step in which training set is analysed (Biswas, Baker and Paquette, 2018). Second
step is classification step in which accuracy of classification rule is estimated. For example: This
method is used by banking industry organizations so that they can identify loan applicants who
have low, medium or high credit risk.
Clustering analysis: It is almost similar to classification but with one differentiation i.e. clusters
are made on the basis of their dependence on similarities of the data items. Different kinds of
clusters have unrelated or dissimilar objectives. Sometimes it is also known as data segmentation
as well because it helps in partitioning huge data sets into clusters as per their similarities. There
are various kinds of clustering methods that are used by organizations as their need and
requirement. For example: if a bank wants to cluster high credit risk customers and lower credit
risk customers on the basis of their age and salary then it can be done in an easy and appropriate
manner.
14

Prediction: this data mining method is used by organizations mostly to predict future on the
basis of past and present trend and data set (Kimmons and Veletsianos, 2018). Prediction is
mostly commonly done by using combination of other mining methods such as pattern matching,
classification, relation and trend analysis. This prediction method can be used by supermarkets.
For example, if a supermarket wants to identify or predict revenue that every item can generate
on the basis of their past sales data.
Pattern tracking or Sequential pattern: It is another kind of most commonly used data mining
method that can be used by organizations so that they can identify patterns that occurs quite
frequently over a certain time period. This can also be used by retail sector organization. For
example, a clothing otherizations uses this method to identify maximum sales of jackets in which
time period and using this method they identify that sales of jacket increase just before winter
season every year which helps in increasing overall profitability of organization during that
particular season.
Decision trees: It is another kind of data mining method which is used by organizations to
classify their items so that important and appropriate decisions can be taken (Patel and et. al.,
2017). For example: this method can be used by Government organizations that issues driving
licence to their citizens, so that they can identify individuals who are 18 or above so that licence
can be issued to them. For this they use this decision tree method in order to classify their
citizens in different age group categories.
Neural network: This method is completely based upon biological neural network. This method
is mostly used by organizations to model relationship between input and output. It is mostly used
by organizations for classification, regression analysis, data processing and many more. It is also
used by organizations for pattern recognition within the input and output in an appropriate and
excellent manner (Kimmons and Veletsianos, 2018).
Outlier Analysis or Anomaly Analysis: This data mining method is mostly used by
organizations for identification of data items that do not comply with expected behaviour or
expected patterns. This identified unexpected data items are known as noise. This method is used
by organizations for various kinds of purposes such as it can be used by banking organizations to
identify credit card fraud detection, detection of fault, intrusion detection and for many more
purposes (Kimmons and Veletsianos, 2018). For example, if a bank wants to detect fault within
their system then they can implement or use those methods so that they can easily identify
15

unexpected data items that are most commonly known as outliers or noise. So that any further
required action which is important to be taken in order to resolve fault issue or problem within
the system can be taken.
These are some of the most commonly used data mining methods that are used by
organizations for different purposes as per their needs and requirements in an appropriate
manner.
Advantages and disadvantages of Weka over Excel
Advantages of Weka over Excel
There are various kinds of advantages because of which Weka is preferred more over
Excel such as: it is an open source, freely available software which is available under GNU
General Public License. It is quite easy to be used because of its graphical user interface. One of
the most important advantage because of which Weka is preferred over Excel is that it is quite
easy to integrate Weka into other packages of java as compared to Excel (Russell and Markov,
2017). There are various kinds of difficulties faced by users if they are required to integrate
Excel into java packages. Another important advantage of Weka because of which it is preferred
more is that it helps in providing packages like range of data prediction, feature selection and
data mining algorithms are integrated. That means that in this only one data format is required in
this. Not only this it can also be used for development of new machine learning schemes. As
compared to Excel it can lead data in many formats such as ARFF, CSV, and C4.5, binary.
Disadvantages of Weka over Excel
Despite of these advantages there are various kinds of limitations or disadvantages that
are associated with Weka because of which it is less preferred over Excel. Some of the main
disadvantages of Weka are: It lacks adequate and proper documentation and has a limitation in
which systems are updated constantly whereas in Excel constant upgrading is not required
(Hussain and et. al., 2018). It has worse connectivity to Excel spreadsheets and to other non-
java-based databases whereas Excel has good connectivity with non- java-based databases. CSV
reader of Weka is not robust as it is in rapid miner. It does not have a facility of saving
parameters for scaling to be applied to future datasets. All kinds of data preparation and
visualization techniques offered by them are not sufficient (Hussain and et. al., 2018). Not doubt
some of them are quite useful but most of the data mining task cannot be completed without
proper and appropriate methods or techniques. Another disadvantage of Weka is scaling. For
16

large datasets running time can be quite long especially if difficult task is being done. Not only
this, many times java also gives out of memory error because of which it can become difficult to
complete the task without any issue or error. Whenever there is requirement of working upon a
large database or data set it would always become a necessary requirement to reduce overall size
of the dataset so that required task can be completed within defined limits and reasonable time
period as well. It does not consist of useful data like test data methods as it is not provided within
the GUI but it can be done in command line interface which is sometimes quite difficult.
CONCLUSION
From above discussion, it concluded that Data mining is consider as a process of finding
the patterns, correlations and anomalies within large data set. In order to predict accurate result
or outcome. It has summarized about the data mining concept for gathering and collecting large
information into single unit. In this report, marketing assistant is focused towards revenue
generations, cut cost while improving the customer relationship. It is mainly focused on the
various risks while adapting the suitable important factor which consider the large volume of
data or information. On the other hand, data mining is exploring and analyzing the large amount
of data which easily discover meaningful pattern, rules and facts. It can be identified the sales
information while primary aim is to predict future result or outcome. In additional, data mining is
an appropriate technique to build machine learning model that power in term of artificial
intelligence.
17

REFERENCES
Book and Journals
Kavakiotis, I., and et. al., 2017. Machine learning and data mining methods in diabetes
research. Computational and structural biotechnology journal. 15. pp.104-116.
Bandaru, S., Ng, A.H. and Deb, K., 2017. Data mining methods for knowledge discovery in
multi-objective optimization: Part A-Survey. Expert Systems with Applications. 70.
pp.139-159.
Biswas, G., Baker, R.S. and Paquette, L., 2018. Data mining methods for assessing self-regulated
learning.
Kimmons, R. and Veletsianos, G., 2018. Public internet data mining methods in instructional
design, educational technology, and online learning research. TechTrends. 62(5).
pp.492-500.
Patel, T.A., and et. al., 2017. Correlating mammographic and pathologic findings in clinical
decision support using natural language processing and data mining
methods. Cancer. 123(1). pp.114-121.
Russell, I. and Markov, Z., 2017, March. An introduction to the Weka data mining system.
In Proceedings of the 2017 ACM SIGCSE Technical Symposium on Computer Science
Education (pp. 742-742).
Hussain, S., and et. al., 2018. Educational data mining and analysis of students’ academic
performance using WEKA. Indonesian Journal of Electrical Engineering and Computer
Science. 9(2). pp.447-459.
18