Sample Research Proposal-Technical Writing-Handouts, Lecture notes of Technical Writing

Download Sample Research Proposal-Technical Writing-Handouts and more Lecture notes Technical Writing in PDF only on Docsity!

DESIGN OF FUZZY EXPERT SYSTEM AND ITS APPLICATION IN

CARDIAC DISEASE DIAGNOSIS USING ECG

Design of Fuzzy Expert System ii

CERTIFICATE OF APPROVAL

This Research Proposal for the Design of Fuzzy Expert System and its Application in Cardiac Disease Diagnosis using ECG is approved for submission to panel.

Design of Fuzzy Expert System iv

TABLE OF CONTENTS

CERTIFICATE OF APPROVAL........................................................................................................ II

ABSTRACT ......................................................................................................................................... III

TABLE OF CONTENTS .....................................................................................................................IV

LIST OF FIGURES .............................................................................................................................. V

LIST OF TABLES ...............................................................................................................................VI

1. INTRODUCTION .......................................................................................................................... 1 1.1 Purpose ............................................................................................................................. 2 1.2 Scope of Research............................................................................................................. 2 1.3 Definitions, Acronyms and Abbreviations ........................................................................ 3 1.4 Overview ........................................................................................................................... 4 1.5 Project Team .................................................................................................................... 4 2. OBJECTIVES OF RESEARCH ...................................................................................................... 5 2.1 Fuzzy Expert Systems ....................................................................................................... 5 2.2 Electrocardiogram ........................................................................................................... 7 2.3 Objectives ....................................................................................................................... 12 **3. LITERATURE REVIEW ............................................................................................................. 15

4. PROJECT PLAN ........................................................................................................................ 20** 4.1 Semester 6th (and summer vacations) ............................................................................ 20 4.2 Semester 7th ................................................................................................................... 20 4.3 Semester 8th ................................................................................................................... 20 4.4 Project Duration ............................................................................................................. 20 **5. CONCLUSION ........................................................................................................................... 21
5. REFERENCES ........................................................................................................................... 22**

Design of Fuzzy Expert System v

LIST OF FIGURES

• [04].................................................................................................................................................. FIGURE 1 IN REALITY MOST OF THE TIME WE NEED SIGNIFICANT KNOWLEDGE NOT PRECISE INFORMATION
• FIGURE 2 GENERIC VIEW OF FES
• FIGURE 3 INITIAL STRUCTURE OF SYSTEM TO BE DESIGNED
• FIGURE 4 COMPONENTS OF ELECTRICAL SYSTEM OF HEART [04]...........................................................
• FIGURE 5 TYPICAL ECG SIGNAL, WITH IMPORTANT COMPONENTS [03]..................................................
• FIGURE 6 DIFFERENT FEATURES OF ECG [08]
• FIGURE 7 LEARNING FUZZY EXPERT SYSTEM
• FIGURE 8 PROBLEM TO CLASSIFY, WHO IS TALL? IS CRISP CLASSIFICATION RIGHT? [04]
• FIGURE 9 CRISP CLASSIFICATION [04]
• FIGURE 10 FUZZY MEMBERSHIP FUNCTION REPRESENTING DEGREE OF MEMBERSHIP [04]
• FIGURE 11 GENERIC DESIGN OF A FUZZY CLASSIFIER [06]
• FIGURE 12 GENERIC DESIGN OF AN ADAPTIVE FUZZY CLASSIFIER [06]

Design of Fuzzy Expert System 1

1. Introduction

This document gives an introduction to the reader to the research proposal of the “Design of Fuzzy Expert System and its Application in Cardiac Disease Diagnosis using ECG” project to be carried out as by the final year students in the Department of Computer & Information Sciences at the Pakistan Institute of Engineering and Applied Sciences (PIEAS) in the year 2007-2008. This section serves as an introduction to this document by defining the problem outline to intended audience for this document. It also outlines the domain specific definitions, acronyms and abbreviations used in this document. Finally it presents the organization of the rest of the document. The name of the project to be designed is “Design of Fuzzy Expert System and its Application in Cardiac Disease Diagnosis using ECG”. The objective of the system to be designed is to use ECG biometrics of a person for the purpose of cardiac disease identification. In the recent years, there is an increase of death rate due to cardiac diseases, early detection of such diseases is crucial because, in some diseases, the effects are reversible. Increased computing power and decreased microchip size has given thrust for implementing powerful diagnostic methods. Early diagnosis is very important in most diseases, like heart diseases, cancer and some others, where later diagnosis may not help in any treatment. In these cases, at initial stages of disease there are very less symptoms and are mostly shown if the patient is observed for long period of time. As there is very low doctor to patient ratio in our country; and that is why it is hard to observe patients for long time by experts but with expert systems it is possible to do that. But these systems mostly are commercial and are much costly. There is need of less complex, easily interpretable and low cost systems. In the area of computer sciences, soft computing is an emerging field which mainly deals with different problems using Fuzzy, Evolutionary and Neural Net based computing. This research project comes under the computational intelligence or soft computing fields. Fuzzy Expert Systems are an open area of

Design of Fuzzy Expert System 2

research; Expert Systems are being designed, since Artificial Intelligence succeeded to present somewhat human knowledge. The application of the designed Fuzzy Expert System will be diagnosis of cardiac diseases using ECG. In this research, the basic aim is to study Fuzzy clustering, Fuzzy classification, Fuzzy Expert Systems and Fuzzy Reasoning and optimizing expert systems, mainly for cardiac disease diagnosis. ECG will be used because it is a non-invasive technique to see how the cardiac electrical system is working. It shows symptoms of diseases which have their effect on cardiac electric- conduction system. The study of different techniques of classification will be used for designing the expert system which will be able to tolerate uncertainty with powerful capabilities of Fuzzy Logic to cope with uncertainty.

1.1 Purpose

This document serves as a control tool for the progress of the research. The objective of this research is to design an effective fuzzy expert system for cardiac disease diagnosis. Cardiac Disease Diagnosis can be classified into following phases:  Preparing Input for FES  Classification of Input  Reasoning of Diagnosis Results

The intended audience to this document is developer (team) and the panel of examiners. This document can be later on used for traceability and for monitoring the progress in research. This document also serves as a starting point for documentation of the research.

1.2 Scope of Research

The scope of research is to design a fuzzy expert system for the diagnosis of cardiac disease. Useful features will be extracted from ECG signals of normal persons and patients of various cardiac disorders. Knowledge acquisition will be done by making a fuzzy rule base extracted from the features of normal persons and patients.

Design of Fuzzy Expert System 4

 FES

Fuzzy Expert System is an expert system which uses fuzzy inference and fuzzy reasoning to achieve a predefined goal  Classification The process of partitioning a feature space onto several regions of decision (output) space  Clustering The process of grouping objects into groups on the basis of similar behavior and characteristics  Diagnosis The identification of a disorder in a patient through physical examination, medical tests and other procedures  ECG Electrocardiogram (ECG) is representation of electrical activity of the heart over time. ECG gives information about the conduction of electric charge through conduction fibers of heart

1.4 Overview

The rest of the document introduces the reader with the components of research to be conducted. The section 2 describes objectives of research. The section 3 gives a brief Literature Review. The section 4 describes project plan.

1.5 Project Team

Student Name Muhammad Usman Akram Project Supervisor Dr. Muhammad Arif Project Coordinator Mr. Syed Muhammad Haroon

Design of Fuzzy Expert System 5

2. Objectives of Research

This research work aims study of FES and related fields. The basic aim is to design and optimization of Fuzzy Expert System. This system will base on Fuzzy Classification/Inference Systems and Reasoning Mechanism.

2.1 Fuzzy Expert Systems

Expert Systems are programs, designed to provide expertise of an expert to other non-experts. Such system attempt to emulate thinking/reasoning of an expert. A fuzzy expert system is an expert system which uses fuzzy rule base, fuzzy inference and fuzzy reasoning. To design expert systems the knowledge and expertise of an expert are gathered and represented in form of rules and fact. Since, beginning of Artificial Intelligence, knowledge acquisition and representation remained difficult tasks; which lead to different approaches for knowledge acquisition, its representation and most important it’s processing. One of the most popular techniques is Neural Networks (where Machine learning is used for knowledge acquisition). But still the knowledge acquisition from human experts remains difficult because of vagueness in natural language, uncertainty of reality and difference in opinion of experts. Most of the techniques are still too rigid and precise to deal with vagueness and uncertainty of reality. So, Lotfi A. Zadeh (University of California at Berkeley) purposed Fuzzy Sets, they were able to represent and work on natural language variable, natural language vagueness and uncertainty of reality. In contrast to other classical methodologies, which require mathematical understanding and precise equations for the system. Fuzzy logic is an alternative technique which gives ease to implement human heuristics and experiences with the system in natural language. In reality most of the time, significant Knowledge is needed, instead of precise information. It will be true to say some time precise information is useless, as shown in figure 1.

Design of Fuzzy Expert System 7

 Defuzzification (Optional): From the degree of truth, finding the crisp output (when crisp needed) is called Defuzzificztion.

Figure 2 Generic View of FES

The designed FES will have a similar structure, as shown in figure 3.

Training DataSet Knowledge Base Classification &Inference

Explanation & Reasoning

Figure 3 Initial Structure of System to be Designed

2.2 Electrocardiogram

Electrocardiogram is representation of the electrical activity of heart. “The electrocardiogram, or ECG / EKG is a surface measurement of the electrical potential generated by electrical activity in cardiac tissue. Current

Design of Fuzzy Expert System 8

flow, in the form of ions, signals contraction of cardiac muscle fibers leading to the heart's pumping action.” [07]

2.2.1 Cardiac background

Heart is the collection of different cardiac muscles. It is the center of the blood circulatory system. Heart is a pump which makes it possible for blood to circulate through out the body in the blood circulatory system. Heart is a electro-mechanical pump, which is controlled by the electric conduction system. Any problem in this conduction system may result as problem in heart functioning.

2.2.1.1 Cardiac Mechanical System

The human heart consists of four chambers. They are:  Right Atrium Right upper chamber is called Right Atrium. Right atria receive deoxygenized blood from body, when it expands. And when it contracts, provides this blood to Right Ventricle.  Left Atrium Left upper chamber is called Left Atrium. Left atria receive oxygenized blood from lungs, when it expands. And when it contracts, provides this blood to Left Ventricle.  Right Ventricle Right lower chamber is called Right Ventricle. It is connected with right atria from which it receives deoxygenized blood, when it expands. And when it contracts, pumps out this blood to lungs.  Left Ventricle Left lower chamber is called Left Ventricle. It is connected with left atria from which it receives oxygenized blood, when is expands. And when it contracts, pumps out this blood to body.

2.2.1.2 Cardiac Electrical System

Mechanical system of the heart works with the support of electrical system. Electric activation moves through conduction fibers. This generates an electric signal which is captured and represented as ECG.

Design of Fuzzy Expert System 10

representation) and mechanical activities. Components of ECG are as they are highlighted in figure 5.

Table 1 Relation of Cardiac Electrical, mechanical activities and ECG Representation [13] Electrical Activity Mechanical Activity ECG Representation Activation of SA Node Atria depolarize Atria contraction Start of P Wave Pause at AV Node Blood flows to ventricles End of P Wave Pulse travels down His Bundle to Bundle Branches

Q wave

Atria repolarize while ventricles depolarize

Atria relax, Ventricles contract pumping blood to lungs and body

R and S wave

Ventricles repolarize Ventricles relax (^) T wave

 P Wave P wave is shows the depolarization of both atria. It is representation of electrical signal that causes the contraction of both atriums.  QRS Complex QRS complex shows the depolarization of ventricle. QRS Complex is representation of electric single that causes the contraction of the ventricles, which is more forceful than that of atriums and involves more cardiac muscles thus resulting in greater ECG deflection.  T Wave T wave represents re polarization of ventricles. This helps their expansion.

Design of Fuzzy Expert System 11

Figure 5 Typical ECG Signal, with important components [03]

 U wave This wave shows up some time due to different reasons.

Main features considered in Diagnosis, prominently shown in figure 6.  Iso-electric line  Voltage (Potential)  Interval  Width  Rates  Regularity  Patterns  Deviation form baseline

Design of Fuzzy Expert System 13

System will be designed or Fuzzy rules will be optimized using Genetic Algorithms. The aim is to design an expert system which will use knowledge base generated by learning just like Neural Networks. The learning methodology may use Knowledge acquisition form data, Adaptive, Neuro system or Evolutionary learning. Study (Understanding and implementation) of following sub systems or techniques will be done for ECG.  Fuzzy Clustering Techniques for Rule-base generation (Knowledge Acquisition)  Fuzzy Classifiers/ Fuzzy Inference Systems (Inference or Decision making)

The above mentioned system may be optimized with further extensions like:  Neuro Fuzzy Techniques  Optimization of implemented techniques will be done using GA’s (Genetic Algorithms)  Designing a Fuzzy Expert System may be using, Evolving / Adaptive Fuzzy System This all implementation and optimization will be done in context of diagnosis of cardiac diseases using ECG.

2.3.1 Relationship between Components

A Fuzzy Expert System can have any of the following designs  Fuzzy Clustering Techniques Preparing initial input for Fuzzy If-Then Rule generation  Fuzzy Classifiers Implementation of Fuzzy If-Then Rules for classification of input  Fuzzy Inference Systems Reasoning using Fuzzy If-Then Rules to explain and justifying the classification

This system design is also shown in figure 7.

Design of Fuzzy Expert System 14

Figure 7 Learning Fuzzy Expert System

A Fuzzy Expert System can be extended to any of the following designs

 Neuro Fuzzy Techniques These techniques are hybrid of Neural Networks and Fuzzy Systems. So, they incorporate learning abilities of Neural Networks and reasoning capabilities of Fuzzy Systems.  Evolving Fuzzy System In Fuzzy Systems rules can be generated by evolution, might be continuous or limited time evolution.  Adaptive Fuzzy System Fuzzy System can be enhanced, if they learn from their own mistakes/ Adaptive Fuzzy Systems adapt to any change in input with very little change in rule base.