International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056

Volume: 02 Issue: 09 | Dec-2015 www.irjet.net p-ISSN: 2395-0072

_________________________________________________________________________________________________

NEURAL NETWORKS IN DATA MINING

S.Sathyamoorthy,
Assistant Professor,
Department of Computer Science/CA/IT,
Bharathiar University Arts and Science College, Gudalur,The Nilgiris,Tamilnadu,India.
-----------------------------------------------***------------------------------------------------
Abstract- Data mining is extracts the knowledge/
information from a large amount of data which stores
in multiple heterogeneous data base. Knowledge
/information are conveying the message through direct
or indirect. There is many technologies available to
data mining practitioners, including Artificial Neural
Networks, Regression, and Decision Trees. Many
practitioners are wary of Neural Networks due to their
black box nature, even though they have proven
themselves in many situations.

Keywords: Data mining, Artificial Neural Network,

Feed forward neural networks.

1. INTRODUCTION:

Data mining is the term used to describe the process of

extracting value from a database. A data-warehouse is a
location where information is stored. The type of data
stored depends largely on the type of industry and the
company. Many companies store every piece of data they
have collected, while others are more ruthless in what
they deem to be “important”. Consider the following 2. ARTIFICIAL NEURAL NETWORKS:
example of a financial institution failing to utilize their
data-warehouse. The development of Information An artificial intelligence is a very general term but
Technology has generated large amount of databases and defining it precisely is very difficult. And the design of an
huge data in various areas. The research in databases and artificially intelligent agent totally depends on the fact
information technology has given rise to an approach to how we define the term ‘Artificial Intelligence’. An
store and manipulate this precious data for further artificial neural network (ANN), often just called a "neural
decision making. Data mining is a process of extraction of network" (NN), is a mathematical model or computational
useful information and patterns from huge data. It is also model based on biological neural networks, in other
called as knowledge discovery process, knowledge mining words, is an emulation of biological neural system. It
from data, knowledge extraction or data /pattern analysis. consists of an interconnected group of artificial neurons
Data mining is a powerful concept for data analysis and and processes information using a connectionist approach
process of discovery interesting pattern from the huge to computation.
amount of data, data stored in various databases such as
data warehouse, world wide web, external sources.

© 2015, IRJET ISO 9001:2008 Certified Journal Page 431

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 09 | Dec-2015 www.irjet.net p-ISSN: 2395-0072
_________________________________________________________________________________________________

2.2 Neural networks versus conventional

computers

Neural networks take a different approach to problem

solving than that of conventional computers. Conventional
computers use an algorithmic approach i.e. the computer
follows a set of instructions in order to solve a problem.
Unless the specific steps that the computer needs to follow
are known the computer cannot solve the problem. That
restricts the problem solving capability of conventional
computers to problems that we already understand and
know how to solve. But computers would be so much
more useful if they could do things that we don't exactly
know how to do. Neural networks process information in
a similar way the human brain does. The network is
Natural neurons receive signals through synapses located composed of a large number of highly interconnected
on the dendrites or membrane of the neuron. When the processing elements (neurons) working in parallel to
signals received are strong enough (surpass a certain solve a specific problem. They cannot be programmed to
threshold), the neuron is activated and emits a signal perform a specific task. The examples must be selected
carefully otherwise useful time is wasted or even worse
though the axon. This signal might be sent to another
the network might be functioning incorrectly. The
synapse, and might activate other neurons.
disadvantage is that because the network finds out how to
solve the problem by itself, its operation can be
2.1 Why use neural networks? unpredictable.

Neural networks, with their remarkable ability to derive Neural networks and conventional algorithmic computers
meaning from complicated or imprecise data, can be used are not in competition but complement each other. There
to extract patterns and detect trends that are too complex are tasks are more suited to an algorithmic approach like
to be noticed by either humans or other computer arithmetic operations and tasks that are more suited to
techniques. A trained neural network can be thought of as neural networks. Even more, a large number of tasks,
an "expert" in the category of information it has been require systems that use a combination of the two
given to analyze. This expert can then be used to provide approaches (normally a conventional computer is used to
projections given new situations of interest and answer supervise the neural network) in order to perform at
"what if“questions. Other advantages include: maximum efficiency.

1. Adaptive learning: An ability to learn how to do 3. NEURAL NETWORKS IN DATA MINING:

tasks based on the data given for training or
initial experience.
2. Self-Organization: An ANN can create its own Neural networks are non-linear statistical data modeling
organization or representation of the information tools. They can be used to model complex relationships
it receives during learning time. between inputs and outputs or to find patterns in data.
3. Real Time Operation: ANN computations may be Using neural networks as Applications of Neural Networks
in Data Mining as a tool, data warehousing firms are
carried out in parallel, and special hardware
devices are being designed and manufactured harvesting information from datasets in the process
which take advantage of this capability. known as data mining. The difference between these data
warehouses and ordinary databases is that there is actual
4. Fault Tolerance via Redundant Information
Coding: Partial destruction of a network leads to manipulation and cross-fertilization of the data helping
the corresponding degradation of performance. users makes more informed decisions.

© 2015, IRJET ISO 9001:2008 Certified Journal Page 432

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 09 | Dec-2015 www.irjet.net p-ISSN: 2395-0072
_________________________________________________________________________________________________

interpretation of the results, data mining process is the

reiteration of the three phases. The data mining based on
neural network is composed by data preparation, rules
extracting and rules assessment three phases, as shown
below:

A. Data Preparation:
Data preparation is to define and process the mining data
to make it fit specific data mining method. Data
preparation is the first important step in the data mining
and plays a decisive role in the entire data mining process.
It mainly includes the following four processes:

1. Data Clustering: Data cleansing is to fill the vacancy

value of the data, eliminate the noise data and correct the
inconsistencies data in the data.
3.1. Feed forward Neural Network:
2. Data Option: Data option is to select the data arrange
One of the simplest feed forward neural networks (FFNN), and row used in this mining.
such as in Figure, consists of three layers: an input layer,
3. Data Pre-processing: Data pre-processing is to
hidden layer and output layer. In each layer there are one
enhanced process the clean data which has been selected.
or more processing elements (PEs). PEs is meant to
simulate the neurons in the brain and this is why they are 4. Data Expression: Data expression is to transform the
often referred to as neurons or nodes. A PE receives data after preprocessing into the form which can be
inputs from either the outside world or the simplified accepted by the data mining algorithm based on neural
process for training a FFNN is as follows: network. The data mining based on neural network can
only handle numerical data, so it is need to transform the
1. Input data is presented to the network and propagated
sign data into numerical data. The simplest method is to
through the network until it reaches the output layer. This
establish a table with one-to-one correspondence between
forward process produces a predicted output.
the sign data and the numerical data. Although there are
2. The predicted output is subtracted from the actual many data types in relational database, but they all
output and an error value for the networks is calculated. basically can be simply come down to sign data, discrete
numerical data and serial numerical data three logical
3. The neural network then uses supervised learning, data types.
which in most cases is back propagation, to train the
network. Back propagation is a learning algorithm for B. Rules extracting:
adjusting the weights. It starts with the weights between
the output layer PE’s and the last hidden layer PE’s and There are many methods to extract rules, in which the
works backwards through the network. most commonly used methods are LRE method, black-box
method, the method of extracting fuzzy rules, the method
4. Once back propagation has finished, the forward of extracting rules from recursive network, the algorithm
process starts again, and this cycle is continued until the
of binary input and output rules, partial rules extracting
error between predicted and actual outputs is minimized.
algorithm and full rules extracting algorithm.
4. DATA MINING PROCESS BASED ON NEURAL
NETWORKS: C. Rules Assessment:

Data mining process can be composed by three main 1. Although the objective of rules assessment depends on
phases: data preparation, data mining, expression and each specific application, but, in general terms, the rules

© 2015, IRJET ISO 9001:2008 Certified Journal Page 433

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 09 | Dec-2015 www.irjet.net p-ISSN: 2395-0072
_________________________________________________________________________________________________

can be assessed in accordance with the following [6] Elements of Artificial Neural Networks: Kishan
objectives: Mehrotra, Chilukuri K. Mohan, Sanjay Ranka.

2. Find the optimal sequence of extracting rules, making it [7]Industry Application of data mining,
obtains the best results in the given data set; http://www.pearsonhighered.com/samplechapter/
0130862711.pdf
3. Test the accuracy of the rules extracted;
[8] David L Olson, Dursun Delen “Advance data
4. Detect how much knowledge in the neural network has mining techniques” springer 2008
not been extracted; Detect the inconsistency between the
extracted rules and the trained neural network.
BIOGRAPHY

S.Sathyamoorthy was born in Erode,

5. Data mining has importance regarding finding the
Tamilnadu, India on 06-05-1981. He
patterns, forecasting, discovery of knowledge etc., in received his M.Sc degree in Information
different business domains. Artificial Neural Networks Science & Management from Kongu Arts
offer qualitative methods for business and economic & Science College, Erode, Tamilnadu,
systems that traditional quantitative tools in statistics and India, in 2004.He is now working as an
econometrics cannot quantify due to the complexity in Assistant Professor in the department of
Computer science/CA/IT at Bharathiar
translating the systems into precise mathematical
University Arts and Science College,
functions. At present, data mining is a new and important
Gudalur, The Nilgiris,Tamilnadu,India .
area of research, and neural network itself is very suitable
for solving the problems of data mining because its
characteristics of good robustness, self-organizing
adaptive, parallel processing, distributed storage and high
degree of fault tolerance.

REFERENCES

[1] IEEE Transactions on Neural Networks; “Data Mining

in a Soft Computing Framework: A Survey”, Authors:
Sushmita Mitra, Sankar K. Pal and Pabitra Mitra. (January
2002, Vol. 13, No. 1)

[2] Using Neural Networks for Data Mining: Mark W.

Craven, Jude W. Shavlik

[3] Berson, “Data Warehousing, Data-Mining & OLAP”,

TMH

[4] Data Mining Techniques: Arjun K. Pujari

[4] Introduction to the theory of Neural Computation:

John Hertz, Anders Krogh, Richard G. Palmer

[5] Bradley, I., Introduction to Neural Networks, Multinet

Systems Pty Ltd 1997.

© 2015, IRJET ISO 9001:2008 Certified Journal Page 434