机器学习驱动的软件故障预测提升：案例推理方法的研究进展

需积分: 0 142 浏览量更新于2024-07-18 收藏 5.59MB PDF 举报

"《增强软件故障预测的机器学习方法：新兴研究与机会》是一本由Ekbal Rashid编著，收录在2018年出版的'系统分析、软件工程与高性能计算'系列（Advances in Systems Analysis, Software Engineering, and High Performance Computing, ASASEHPC）书籍中的作品。该书由国际IGI全球出版社发行，位于美国赫尔希，探讨了如何利用机器学习技术改进软件质量预测，特别是通过案例推理（Case-Based Reasoning, CBR）这一方法。 CBR是一种与众不同的AI技术，其核心在于它依赖于已有的经验知识库，而非复杂的领域特定规则。相比于其他AI技术，CBR的优势在于它更容易获取和理解，因为它是以实例形式存在的，而不是抽象的规则。在软件故障预测领域，作者关注的是如何通过案例学习来解决实际问题，这种方法能够有效地从过去的错误或成功案例中提取模式，从而提高预测准确性。随着近年来机器学习的快速发展，特别是在大数据和人工智能的推动下，CBR作为软件故障预测的一种创新手段，为研究者提供了新的研究方向和实践机遇。书中可能包含了对不同机器学习算法的应用分析，如监督学习、无监督学习以及深度学习等，这些方法如何在实际软件开发环境中提升故障检测和预防能力。此外，书中还可能涵盖了如何收集、整理和评估案例数据，如何构建有效的案例库，以及如何通过特征选择和模型训练来优化预测性能。书中可能会提供一些实证研究结果，展示了CBR在软件故障预测中的有效性，并可能对未来的研究方向和挑战进行了深入探讨。《增强软件故障预测的机器学习方法》是一本具有实用价值和技术深度的著作，对于软件工程专业人士、研究人员以及机器学习爱好者来说，是一份宝贵的学习资源和参考文献。"

Chapter 1

DOI: 10.4018/978-1-5225-3185-2.ch001

ABSTRACT

The basic aim of software quality (fault) is to identify error prone tasks as the

cost can be minimized with advance knowledge about the errors and this early

treatment of error will enhance the software quality. Today quality is critical for

survival and success. Customers demand it. Software is now a global business

and organizations will not succeed in this global market unless they produce

quality products/services and unless the organizations are also perceived by

customers to produce quality products and services. I need measures of the

characteristics and quality parameters of the development process and its

stages, as well as metrics and models to ensure that the development process

is under control and moving toward the product’s quality objectives.

INTRODUCTION

Since quality is the first step for improvement in anything, to achieve an

improved state of affairs, quality must be precisely defined and appropriately

measured. But very often in quality engineering and management, the term

quality is misunderstood on account of its ambiguous characteristics. The

ambiguity aspect of quality owes its origin from the following (Kan, 2012);

• Quality being multi-dimensional with dimensions of entity of interest,

the viewpoint on that entity and the quality attributes of that entity gets

interpreted diﬀerently depending upon the situation and interpreter.

Software Quality

• On the basis of conceptualizing quality, it may be referred either in its

broadest sense or in its precise connotation.

• Being an element of our everyday language, quality may be interpreted

diﬀerently depending upon the uses – popular or professional.

However, as the focus of this book will be around the quality of software

which is a sort of the professional use of the concept quality as against the

popular use of the same, the third point mentioned above duly makes it very

clear how quality in popular view is quite different from quality in professional

view and hence it draws our attention to take care.

QUALITY IN POPULAR CONNOTATION

Quality in popular connotation highlights the ordinal dimension of the concept

discarding the cardinal one, as per which;

• Quality is an elusive attribute that can be felt, talked about and

compared but cannot be quantiﬁed or measured.

• Quality means opulence, class and taste, i.e., expensive, elaborate and

more complex.

In the former sense, if I use two different brands of a product or service,

I can feel which one is better, but I won’t be possible to express by how

much one is better or worse than the other. In fact, under this viewpoint,

psychological inclination will be there towards one brand among a group of

brands depending upon the liking of the user towards that particular brand.

Thus, this assessment will be subjective as it varies from person to person.

Accordingly, generalizing whether a brand is qualitative or not is exceedingly

tough and rewarding the brand as per the quality is rather tougher. In the

latter sense, quality is limited to an undersized group of elegant products or

services with sophisticated functionality and items that have a touch of class.

QUALITY IN PROFESSIONAL CONNOTATION

Quality in popular connotation, as I experience, invites delusion and

imprecision. Still quality has been and is being treated as the focal point

for any firm to have an edge over other competitors. However, the fallacies

Software Quality

of the popular connotation of quality no way help the quality up-gradation

efforts in the industries. In fact, to develop strategic quality improvement

efforts for the betterment of the organizations, quality must be defined in a

very professional way. To that extent, the definitions given by Crosby (1979),

and Juran and Gryna (1970) seem to be appropriate and hence being used by

many quality professionals. While Crosby defines quality as ‘conformance

to requirements’ Juran and Gryna defines it as ‘fitness for use.’

Both the definitions give importance to customers’ requirements and

expectations. While Crosby’s view point clusters around the producer’s

assessment of consumers’ requirements and developing the product or service

accordingly, the view point of Juran and Gryna focus on consumers’ assessment

of performance of the product or service. Thus, it is very essential to develop

a product or service exactly in line with the customers’ requirements and

modify if required in further phases analyzing the post-purchase reaction of the

customers. From Customers’ view point quality is the perceived value of the

product or service. In fact, before purchasing a product or service, a customer

has a perceived value on that. After purchasing that he/she may derive a level

of satisfaction which may be exactly the same as his/her perceived value or

more or less. If the realized value happens to be the same as the perceived

value, the customer does not find any gap or the producer does not allow

finding a gap. That leads to tag the product a quality product. Contrary to

this, if the realized value happens to be more than the perceived value, the

customer becomes over satisfied and becomes addicted to that brand of the

product or service. In such a situation, he/she may refer his/her relatives and

friends in favor of the product or service. Accordingly, the performance of that

product or service will get due recognition in the market. If the realized value

happens to be less than the perceived value, not only the concerned customer

will be unhappy, but due to bad word of mouth the market performance of

the product or service will also be affected.

While customer buys a product or service he/she compares the assurances

given by the producer with his/her expectations for that product or service. In

fact, the sale of product depends upon the level of assurances for the quality

and the extent to which the assurances are meeting the expectations of the

customers. Thus, the customers can appropriately say about the quality of any

product or service. To perform well in the market and maintain sustainability

pertaining to the market share a product or service has to go for marketing

research at development stage to know the customers’ expectations or

requirements and on post-purchase stage to know the reaction of the customers

after using a product or service. The research findings of these researches will

Software Quality

help the producer to produce quality output and make necessary changes in

the subsequent phases of production depending upon the customers’ reactions.

This way of producing quality product or service and maintaining the same

in the long run will also be applicable to software products.

THE ROLE OF END USERS AND PRIME CUSTOMER

The role of customers and end users is very important in determining the

quality parameters of a product. The end-user might define quality as something

which is user-friendly and bug-free. If the customer is in perfect position to

tell about his expectations and requirements for a product, then it becomes

easier to define the quality parameters of that product in quantitative terms. On

the other hand, if the customer is not specific about the requirements (mainly

in the case of a software product) the quality parameters may remain vague.

The role of the customer, as it relates to quality, can never be overstated.

From a customer’s point of view, quality is the customer’s perceived value

of the product he or she has purchased, based on a variety of variables such

as price, performance, reliability, and satisfaction. The customers are in a

perfect position to tell you about the quality, because that’s all they’re really

buying. They’re not buying a product. They’re buying the assurances that their

expectations for that product will be met. And you haven’t really got anything

else to sell them, but those assurances and quality. From a concept’s high-level

definition to a product’s operational definition, many steps are involved, each

of which may be vulnerable to the shortcomings. For example, to achieve the

state of conformance to requirements, the customers’ requirements must be

first gathered and analyzed, specifications to meet those requirements must be

produced, and the product must be developed and manufactured accordingly.

In each phase of the process, errors can occur that will affect the quality of

the finished product. The requirements may be erroneous (this is especially

the case for software development), the development and manufacturing

process may be subject to variables that induce defects, and so forth. From

the customer’s perspective, satisfaction after the purchase of the product is

the ultimate validation that the product conforms to the requirements and is

fit to use. From the producer’s perspective, once requirements are specified,

developing and producing the product in accordance with the specifications

is the path to achieving quality. Usually, for product quality, the lack of

defects and good reliability are the most basic measures. Because of the two

perspectives on quality (customer satisfaction as the ultimate validation of

Software Quality

quality and producer’s adherence to requirements to achieve quality), the

de facto definition of quality consists of two levels. The first is the intrinsic

product quality, often operationally limited to the product defect rate and

reliability. This narrow definition is referred to as the “small q” (q for quality).

The broader definition of quality includes product quality, process quality,

and customer satisfaction, and it is referred to as the “big Q.” This two-

level approach to the definition of quality is being used in many industries,

including the automobile industry, the computer industry (both software and

hardware), and the consumer electronics industry (Kan, 2012).

WHAT IS SOFTWARE QUALITY?

Software quality is defined as the quality that ensures customer satisfaction

by offering all the customer deliverables on performance, standards and ease

of operations (Jawadekar, 2006). Software quality can also be defined as

conformance to explicitly stated and implicitly stated functional requirements.

The explicitly stated functional requirements can be derived from the

requirements stated by the customer which are generally documented in some

form. However, implicit requirements are requirements which are not stated

explicitly but are intended. In software, the simple sense of product quality

is commonly known as a free from “bugs” in the product. It is also the most

basic sense of conformance to requirements, because if the software contains

too many functional faults, the basic requirement of providing the desired

function is not met. This definition is usually expressed in two ways: defect

rate (e.g., number of defects per million lines of source code, per function

point, or other unit) and reliability (e.g., number of failures per n hours of

operation, mean time to failure, or the probability of failure-free operation

at a specified time). Customer satisfaction is usually measured by percent

satisfied or unsatisfied (neutral and dissatisfied) from customer satisfaction

surveys. To reduce bias, techniques such as blind surveys (the interviewer

does not know the customer and the customer does not know the company

the interviewer represents) are usually used. In addition to overall customer

satisfaction with the software product, satisfaction towards specific attributes

is also gauged. For example, the higher the functional complexity of the

software, the harder it becomes to achieve maintainability. Depending on the

type of software and customers, different weighting factors are needed for

different quality attributes. For large customers with sophisticated networks and

real-time processing, performance and reliability may be the most important

剩余143页未读，继续阅读

WindStand

粉丝: 35
资源: 367

机器学习驱动的软件故障预测提升：案例推理方法的研究进展

"大规模游戏社交网络节点相似性算法及其在云上深度学习训练中的应用

"杨教授经典著作：迁移学习的利器，打造可靠、稳健的机器学习系统

"机器学习、深度学习画图PPT模板-ML Visuals By dair.ai

Enhancing-the-performance-of-a-tensioned-meta_2018_Nuclear-Instruments-and-M

Enhancing the Discriminative Feature Learning for Visible-Thermal Cross-Modality

Enhancing stock market prediction with extended coupled hidden Markov model over multi-sourced data.

Towards-enhancing-grid-reliability-A-framework-to-forecast-electricity-demand-growth

ExtremeLearningMachine资源共享-Towards-enhancing-centroid-classifier-for-text-classification_2013_Neurocomp.pdf

matlab加密代码-enhancing-security-using-digital-image-processing:通过对发送方和接收方

Enhancing physical-layer security via big-dataaided hybrid relay selection

最新资源