ARTIFICIAL NEURAL NETWORK

A
Seminar Report
Submitted
In partial fulfilment
For the award of the Degree of
Bachelor of Technology
In
Electrical Engineering

Submitted to :

Submitted By:

Mr. Dwarka Prasad

Praveen Sharma

(Associate Prof.)

Roll No. 11ELDEE301

(HOD E.E.)

Department of Electrical Engineering

Laxmi Devi Institute of Engineering & Technology Chikani,
Alwar-301028 (Rajasthan)
March-2015

DECLARATION
I, am Praveen Sharma student of Electrical Engineering of Laxmi Devi Institute of
Engineering & Technology, Alwar hereby declare that the all information written in this
seminar report file are true.
The seminar entitled assigned to me by ARTIFICIAL NEURAL NETWORK during my
seminar for the partial fulfilment from Rajasthan Technical University, Kota is the original
work done by me and the information provided in the study is authentic to the best of my
knowledge.

(PRAVEEN SHARMA)

ACKNOWLEDGEMENT
I, Praveen Sharma student of B.Tech. Final Year EE, am highly grateful to Mr.
DWARKA PRASAD, HOD EE Dept., for his untiring help and encouragement during the
course of seminar titled ARTIFICIAL NEURAL NETWORK. I highly acknowledge, with
deepest sense of gratitude and indebtedness, his great coordination and support throughout
the course of work.
I wish to express my deep sense of gratitude to all the faculty members of Electrical
Engineering Dept. for their guidance and useful suggestions, which helped me in completing
the seminar work, in time

PRAVEEN SHARMA
8th Sem, EE,
11ELDEE301

TABLE OF CONTENTS
Sr. No.

Topic

Page No.

Candidate Declaration

ii

Acknowledgment

iii

Contents

iv-vi

List of Figures

vii

Abstract

viii

CHAPTER 1. Artificial Neural Networks

1. Introduction

2. Review

3. Development of an ANN Model

3.1 Variable Selection

3.2 Formation of Training, Testing and Validation Sets

3.3 Neural Network Architecture

3.4 Model Building

4. Architecture of Neural Networks

5. Supervised and Unsupervised Learning

6. Further Concepts on Hidden Layers

CHAPTER 2. Artificial Neural Networks For Pattern Recognition

12

1. Introduction

12

2. Patterns and pattern recognition tasks

12

2.1 Notion of intelligence

12

2.2 Pattern recognition tasks

13

2.3 Methods for pattern recognition tasks

15

CHAPTER 3. Basics of artificial neural networks

18

1.

Characteristics of biological neural networks

18

2.

Artificial neural networks terminology

20

2.1 Processing unit:

20

2.2 Interconnections :

21

2.3 Operations

21

2.4 Update:

21

3.

22

Models of neurons

3.1 Mc-Culloch-Pitts model:

22

3.2 Perceptron

23

3.3 Adaline:

24

3.4 Topology

24

3.5 Activation and synaptic dynamics

26

CHAPTER 4. Artificial Neural Networks for Beginners

29

1. Introduction

29

2. Networks

29

3. Artificial neural networks

30

4. The Backpropagation Algorithm

32

CHAPTER 5. Basic Learning Principles of Artificial Neural Networks

37

1. Introduction

37

2. Basic Structure of the BPNN Model

38

3. Learning Process of the BPNN Algorithm

39

4. Weight Update Formulae of the BPNN Algorithm

40

5 Conclusions

47

CHAPTER 6. Introduction to Artificial Neural Network Methods

48

1.

Introduction

48

2.

Problem consideration

49

3.

Basic concepts of ANNs

49

4.

Learning by ANN

54

5.

Applications

57

REFRENCES

58

LIST OF FIGURES
FIG. NO.

FIG. NAME

PG. NO.

Figure 1

Schematic representation of neural network

Figure 2

Mathematical representation of neural network

Figure 3

A learning cycle in the ANN model

Figure 4

Schematic drawing of a typical neuron or nerve cell.

17

Figure 5

MP model of a neuron

21

Figure 6

Some nonlinear functions. (a) Binary, (b) ramp and (c) sigmoid

22

Figure 7

Rosenblatt's model of a neuron.

22

Figure 8

Widrow's adaline model of a neuron.

23

Figure 9

Some basic structures of the Artificial Neural Networks.

24

Figure 10

Natural neurons (artists conception).

29

Figure 11

An artificial neuron

29

Figure 12

The basic architecture of neural network

36

Figure 13

Neural network as a black-box featuring the non-linear relationship 47

Figure 14 .

Comparison between the biological and artificial neuron.

50

Figure 15

Three different transfer functions

50

Figure 16

One-layer (left) and two-layer (right) ANNs

52

Figure 17

Order for correction of weights in the error back-propagation

55

ABSTRACT
7

Artificial Neural Networks (ANNs) are non-linear mapping structures based on the
function of the human brain. They are powerful tools for modelling, especially when the
underlying data relationship is unknown. ANNs can identify and learn correlated patterns
between input data sets and corresponding target values. After training, ANNs can be used to
predict the outcome of new independent input data. ANNs imitate the learning process of the
human brain and can process problems involving non-linear and complex data even if the
data are imprecise and noisy. Thus they are ideally suited for the modeling of agricultural
data which are known to be complex and often non-linear. ANNs has great capacity in
predictive modeling i.e., all the characters describing the unknown situation can be presented
to the trained ANNs, and then prediction of agricultural systems is guaranteed.
An ANN is a computational structure that is inspired by observed process in natural networks
of biological neurons in the brain. It consists of simple computational units called neurons,
which are highly interconnected. ANNs have become the focus of much attention, largely
because of their wide range of applicability and the ease with which they can treat
complicated problems. ANNs are parallel computational models comprised of densely
interconnected adaptive processing units. These networks are fine-grained parallel
implementations of nonlinear static or dynamic systems. A very important feature of these
networks is their adaptive nature, where learning by example replaces programming in
solving problems. This feature makes such computational models very appealing in
application domains where one has little or incomplete understanding of the problem to be
solved but where training data is readily available. ANNs are now being increasingly
recognized in the area of classification and prediction, where regression model and other
related statistical techniques have traditionally been employed. The most widely used
learning algorithm in an ANN is the Backpropagation algorithm. There are various types of
ANNs like Multilayered Perceptron, Radial Basis Function and Kohonen networks. These
networks are neural in the sense that they may have been inspired by neuroscience
but not necessarily because they are faithful models of biological neural or cognitive
phenomena. In fact majority of the network are more closely related to traditional
mathematical and/or statistical models such as non-parametric pattern classifiers, clustering
algorithms, nonlinear filters, and statistical regression models than they are to neurobiology
models.