CASE BASED REASONING IN MEDICAL DIAGNOSIS

Case Based Reasoning In Medical Diagnosis

Abstract

Case-based reasoning (CBR) is one of the most well liked proposition methods in health domains because it is so straightforward to request, has no likelihood of over fitting, and presents a good interpretation for the output. However, it has a critical limitation - its proposition presentation is usually smaller than other AI methods like artificial neural systems (ANN). In alignment to get unquestionable outcomes from CBR, productive retrieval and equivalent of helpful former situations for the difficulty is absolutely crucial, but it is still a contentious topic to conceive a good equivalent and retrieval means for CBR systems. In this study, we suggest a innovative set about to enhance the proposition presentation of CBR. Our proposal is the simultaneous optimization of characteristic weights, example assortment, and the number of friends that blend utilising genetic algorithms (GA). Our form advances the proposition presentation in three modes - (1) assessing likeness between situations more unquestionably by contemplating relation significance of each characteristic, (2) eradicating ineffective or mistaken quotation situations, and (3) blending some alike situations comprise important patterns. To validate the utility of our form, this study directed it to a real-world case for assessing cytological characteristics drawn from exactly from a digital scan of breast fine needle aspirate (FNA) slides. Experimental outcomes displayed that the proposition correctness of accepted CBR may be advanced considerably by utilising our model. We furthermore discovered that our suggested form outperformed all the other optimized forms for CBR utilising GA.

Table of Contents

Abstract2

CHAPTER 17

Introduction7

Back ground of Study8

Research Question8

Purpose Of Study9

Significance10

Chapter 211

Literature Reviewa11

Case-based reasoning and optimization models11

Feature selection and weighting approaches in CBR13

Instance selection approaches15

Optimization of the number of neighbors that combine16

Simultaneous optimization approaches19

Genetic algorithms for optimizing factors in case-based reasoning21

Global optimization of feature weights, instance selection, and the number of neighbors that combine using genetic algorithms23

Case-based reasoning27

Medical case-based reasoning systems29

System architecture32

Case retrieval34

Case adaptation36

Revise case37

Save case38

The knowledge creating phase38

The knowledge inferring phase40

Case Representation:46

Case Indexing Process:46

CBR-BASED SYSTEM FOR DIAGNOSIS OF CANCER DISEASES49

CBR-BASED SYSTEM FOR DIAGNOSIS OF HEART DISEASES50

Adaptations of chosen therapies to laboratory findings51

Resistance information52

Prototypes54

Selection of a prototype tree55

Prototype generation strategies56

Chapter 364

Methadology64

The research design and experiments64

Chapter 470

Results70

Experimental results70

Chapter 573

Conclusions73

References76

CHAPTER 1

Introduction

Case-based reasoning (CBR) is a problem-solving method that is alike to the conclusion producing method that human beings use in numerous real-world applications. It often displays important pledge for advancing the effectiveness of convoluted and unstructured conclusion making. In idea, there is no likelihood of over fitting in CBR because it values exact information of before skilled difficulties other than their generalized patterns. CBR is sustained in an up-to-date state because the case-base is revised in real-time, which is a very significant characteristic for the real-world application. Also, it can interpret why it presents a answer by giving alike vintage cases. Consequently, it has been directed to diverse problem-solving localities encompassing technology, investment, trading, and medical diagnosis. In specific, CBR is very befitting for medical submissions because the characteristics of CBR fit to medical domains very ...