predicting fault-prone modules with case-based reasoning

2
( ) Engineering Information Abstracts Part II 170 process is recognized to be among the highest quality and highest reliability software in operation in the world. In En- glish EI Order Number: EIP98014022055 Keywords: Software engineering; Computer software selection and evaluation; Quality assurance; Standards; Space applica- tions Title: GUIDING REENGINEERING WITH THE OPERATIO- NAL PROFILE () Author s : Cusick, James; Fine, Max Corporate Source: AT&T, Bridgewater, NJ, USA Conference Title: Proceedings of the 1997 8th International Symposium on Software Reliability Engineering, ISSRE Con- ference Location: Albuquerque, NM, USA Conference Date: 19971102-19971105 Source: Proceedings of the International Symposium on Soft- ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc, Los Alamitos, CA, USA,97TB100170. p 111 CODEN: PSS- RFV ISSN: 1071-9458 Publication Year: 1997 Abstract: Results of applied Software Reliability Engineering Ž . SRE processes and tools on two teleconference support systems are documented. Beginning with the initial evaluation of SRE technology for AT&T’s TeleConference Service De- velopment in 1992, this presentation covers the eventual use of SRE in many phases of software development for two separate systems. Introduction of SRE into the system test and production support processes of a legacy system, in con- junction with other development process improvements, brought early rewards in quality and timely delivery. SRE provided a gauge by which quality efforts could be measured. Ramp up time and costs, as well as both successes and failures in the implementation of SRE are discussed. Next, presenta- tion of preliminary findings regarding the use of SRE during the analysis, design, and integration testing phases of a next generation ClientrServer conferencing system are discussed. Reuse of an Operational Profile and SRE tools originally developed and deployed for the legacy system to guide design on a new mixed platform ClientrServer architecture incor- porating a Windows interface and Unix application server is explained. Lastly, a discussion of the appropriateness of SRE methods for both traditional functional decomposition tech- niques and Object-Oriented analysis and design is offered. Ž . Author abstract In English EI Order Number: EIP98014022054 Keywords: Computer aided software engineering; Computer software selection and evaluation; Quality assurance; Telecon- ferencing; Computer architecture; UNIX; Object oriented Ž . programming; Interfaces computer ; Computer operating sys- tems Title: DETECTION OF RESPONSE TIME FAILURES OF REAL-TIME SOFTWARE () Author s : Pekilis, B.R.; Seviora, R.E. Corporate Source: Univ of Waterloo, Waterloo, Ont, Can Conference Title: Proceedings of the 1997 8th International Symposium on Software Reliability Engineering, ISSRE Con- ference Location: Albuquerque, NM, USA Conference Date: 19971102-19971105 Source: Proceedings of the International Symposium on Soft- ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc, Los Alamitos, CA, USA,97TB100170. p 38-47 CODEN: PSS- RFV ISSN: 1071-9458 Publication Year: 1997 Abstract: In real-time software, a correct output which is not produced within a specified response time interval constitutes a failure. The detection of response times failures is done by a separate unit which observes the inputs and outputs of the target software. Automatic detection of such failures is com- plicated by state dependencies which require the unit to track a target’s state as well as the elapsed times between specified stimulus and response pairs. A novel, black-box approach is described for detecting response time failures and quality of service degradations of session-oriented, real-time software. 16 Refs. In English EI Order Number: EIP98014022047 Keywords: Software engineering; Computer software selection and evaluation; Quality assurance; Real time systems; Re- Ž . sponse time computer systems ; Computer system recovery; Computer hardware description languages Title: PREDICTING FAULT-PRONE MODULES WITH CASE-BASED REASONING () Author s : Khoshgoftaar, Taghi M.; Ganesan, K.; Allen, Ed- ward B.; Ross, Fletcher D.; Munikoti, Rama; Goel, Nishith; Nandi, Amit Corporate Source: Florida Atlantic Univ, Boca Raton, FL, USA Conference Title: Proceedings of the 1997 8th International Symposium on Software Reliability Engineering, ISSRE Con- ference Location: Albuquerque, NM, USA Conference Date: 19971102-19971105 Source: Proceedings of the International Symposium on Soft- ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc, Los Alamitos, CA, USA,97TB100170. p 27-35 CODEN: PSS- RFV ISSN: 1071-9458 Publication Year: 1997 Abstract: Software quality classification models seek to predict quality factors such as whether a module will be fault-prone, Ž . or not. Case-based reasoning CBR is a modeling technique that seeks to answer new questions by identifying similar ‘cases’ from the past. When applied to software reliability, the working hypothesis of our approach is this: a module currently under development is probably fault-prone if a module with similar product and process attributes in an earlier release was fault-prone. The contribution of this paper is application of case-based reasoning to software quality modeling. To the best of our knowledge, this is the first time that case-based reasoning has been used to identify fault-prone modules. A case study illustrates our approach and provides evidence that case-based reasoning can be the basis for useful software quality classification models that are competitive with dis- criminant models. The case study revisits data from a previ- ously published nonparametric discriminant analysis study. The Type II misclassification rate of the CBR model was

Upload: taghi-m-khoshgoftaar

Post on 17-Sep-2016

218 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: Predicting fault-prone modules with case-based reasoning

( )Engineering Information Abstracts Part II170

process is recognized to be among the highest quality andhighest reliability software in operation in the world. In En-glish EI Order Number: EIP98014022055Keywords: Software engineering; Computer software selectionand evaluation; Quality assurance; Standards; Space applica-tions

Title: GUIDING REENGINEERING WITH THE OPERATIO-NAL PROFILE

( )Author s : Cusick, James; Fine, MaxCorporate Source: AT&T, Bridgewater, NJ, USAConference Title: Proceedings of the 1997 8th InternationalSymposium on Software Reliability Engineering, ISSRE Con-ference Location: Albuquerque, NM, USA Conference Date:19971102-19971105Source: Proceedings of the International Symposium on Soft-ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc,Los Alamitos, CA, USA,97TB100170. p 111 CODEN: PSS-RFV ISSN: 1071-9458Publication Year: 1997Abstract: Results of applied Software Reliability EngineeringŽ .SRE processes and tools on two teleconference supportsystems are documented. Beginning with the initial evaluationof SRE technology for AT&T’s TeleConference Service De-velopment in 1992, this presentation covers the eventual useof SRE in many phases of software development for twoseparate systems. Introduction of SRE into the system testand production support processes of a legacy system, in con-junction with other development process improvements,brought early rewards in quality and timely delivery. SREprovided a gauge by which quality efforts could be measured.Ramp up time and costs, as well as both successes and failuresin the implementation of SRE are discussed. Next, presenta-tion of preliminary findings regarding the use of SRE duringthe analysis, design, and integration testing phases of a nextgeneration ClientrServer conferencing system are discussed.Reuse of an Operational Profile and SRE tools originallydeveloped and deployed for the legacy system to guide designon a new mixed platform ClientrServer architecture incor-porating a Windows interface and Unix application server isexplained. Lastly, a discussion of the appropriateness of SREmethods for both traditional functional decomposition tech-niques and Object-Oriented analysis and design is offered.Ž .Author abstract In English EI Order Number:EIP98014022054Keywords: Computer aided software engineering; Computersoftware selection and evaluation; Quality assurance; Telecon-ferencing; Computer architecture; UNIX; Object oriented

Ž .programming; Interfaces computer ; Computer operating sys-tems

Title: DETECTION OF RESPONSE TIME FAILURES OFREAL-TIME SOFTWARE

( )Author s : Pekilis, B.R.; Seviora, R.E.Corporate Source: Univ of Waterloo, Waterloo, Ont, CanConference Title: Proceedings of the 1997 8th InternationalSymposium on Software Reliability Engineering, ISSRE Con-

ference Location: Albuquerque, NM, USA Conference Date:19971102-19971105Source: Proceedings of the International Symposium on Soft-ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc,Los Alamitos, CA, USA,97TB100170. p 38-47 CODEN: PSS-RFV ISSN: 1071-9458Publication Year: 1997Abstract: In real-time software, a correct output which is notproduced within a specified response time interval constitutesa failure. The detection of response times failures is done by aseparate unit which observes the inputs and outputs of thetarget software. Automatic detection of such failures is com-plicated by state dependencies which require the unit to tracka target’s state as well as the elapsed times between specifiedstimulus and response pairs. A novel, black-box approach isdescribed for detecting response time failures and quality ofservice degradations of session-oriented, real-time software.16 Refs. In English EI Order Number: EIP98014022047Keywords: Software engineering; Computer software selectionand evaluation; Quality assurance; Real time systems; Re-

Ž .sponse time computer systems ; Computer system recovery;Computer hardware description languages

Title: PREDICTING FAULT-PRONE MODULES WITHCASE-BASED REASONING

( )Author s : Khoshgoftaar, Taghi M.; Ganesan, K.; Allen, Ed-ward B.; Ross, Fletcher D.; Munikoti, Rama; Goel, Nishith;Nandi, AmitCorporate Source: Florida Atlantic Univ, Boca Raton, FL,USAConference Title: Proceedings of the 1997 8th InternationalSymposium on Software Reliability Engineering, ISSRE Con-ference Location: Albuquerque, NM, USA Conference Date:19971102-19971105Source: Proceedings of the International Symposium on Soft-ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc,Los Alamitos, CA, USA,97TB100170. p 27-35 CODEN: PSS-RFV ISSN: 1071-9458Publication Year: 1997Abstract: Software quality classification models seek to predictquality factors such as whether a module will be fault-prone,

Ž .or not. Case-based reasoning CBR is a modeling techniquethat seeks to answer new questions by identifying similar‘cases’ from the past. When applied to software reliability, theworking hypothesis of our approach is this: a module currentlyunder development is probably fault-prone if a module withsimilar product and process attributes in an earlier releasewas fault-prone. The contribution of this paper is applicationof case-based reasoning to software quality modeling. To thebest of our knowledge, this is the first time that case-basedreasoning has been used to identify fault-prone modules. Acase study illustrates our approach and provides evidence thatcase-based reasoning can be the basis for useful softwarequality classification models that are competitive with dis-criminant models. The case study revisits data from a previ-ously published nonparametric discriminant analysis study.The Type II misclassification rate of the CBR model was

Page 2: Predicting fault-prone modules with case-based reasoning

( )Engineering Information Abstracts Part II 171

substantially better than that of the discriminant model. Al-though the Type I misclassification rate was slightly greaterand the overall misclassification rate was only slightly less, theCBR model was preferred when costs of misclassification were

Ž .considered. Author abstract 19 Refs. In English EI OrderNumber: EIP98014022046Keywords: Software engineering; Quality assurance; Artificialintelligence; Computer software selection and evaluation

Title: STATISTICAL QUALITY CONTROL: HOW TO DE-TECT UNRELIABLE SOFTWARE COMPONENTS

( )Author s : Stieber, Harald A.Corporate Source: Polytechnic Univ of Nuremberg, Nurem-berg, GerConference Title: Proceedings of the 1997 8th InternationalSymposium on Software Reliability Engineering, ISSRE Con-ference Location: Albuquerque, NM, USA Conference Date:19971102-19971105Source: Proceedings of the International Symposium on Soft-ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc,Los Alamitos, CA, USA,97TB100170. p 8-12 CODEN: PSS-RFV ISSN: 1071-9458Publication Year: 1997

ŽAbstract: If ‘classical’ testing-strategies are used no usage-.testing the application of software reliability growth models

may be difficult and reliability predictions can be misleading.Nevertheless statistical methods can be successfully applied tofailure data. This paper presents an approach which allows thedetection of unreliable software components and the compar-ison of the reliability of different software versions - even iftesting is done in a classical manner. A simple to use graphicalmethod - which is mainly based on the sequential test of Waldis described. The methodology was successfully applied to a

Ž .software system for tax consultants. Author abstract 10 Refs.In English EI Order Number: EIP98014022044Keywords: Software engineering; Computer software selectionand evaluation; Statistical methods; Quality control

Title: PROCEEDINGS OF THE 1997 8TH INTERNATIONALSYMPOSIUM ON SOFTWARE RELIABILITY ENGINEER-ING, ISSRE

( ) Ž .Author s : Anon Ed.Conference Title: Proceedings of the 1997 8th InternationalSymposium on Software Reliability Engineering, ISSRE Con-ference Location: Albuquerque, NM, USA Conference Date:19971102-19971105Source: Proceedings of the International Symposium on Soft-ware Reliability Engineering, ISSRE 1997. IEEE Comp Soc,Los Alamitos, CA, USA,97TB100170. 418p CODEN: PSSRFVISSN: 1071-9458Publication Year: 1997Abstract: The proceedings contains 41 papers from the 8thInternational Symposium on Software Reliability Engineering.Topics discussed include: detection of unreliable softwarecomponents; software reliability problems; fault-prone mod-ules with case-based reasoning; detection of response timefailures of real-time software; hierarchical supervisor for auto-matic detection of software failures; controlling test coverage;

software reliability in weapon systems; reversible computing;large-scale software testing; software quality classificationtrees; computer supported code reviews; confidence-based re-liability; statistical coverage estimation; common-mode fail-ures; reliability allocation; and software triggers. In English EIOrder Number: EIP98014021817Keywords: Software engineering; Computer software selectionand evaluation; Quality assurance; Neural networks; Backpro-pagation; Real time systems; Computer system recovery; Re-

Ž . Ž .sponse time computer systems ; Codes symbols ; Programdebugging

Title: FIELDBUS - KEY TO DIAGNOSTICS AND EFFI-CIENT INSTRUMENT MANAGEMENT

( )Author s : Berge, JonasCorporate Source: Smar, Singapore, SingaporeConference Title: Proceedings of the 1997 ISA TECHrEXPO

Ž .Technology Update. Part 1r3 of 5 Conference Location:Anaheim, CA, USA Conference Date: 19971007-19971009Source: Networking, Industrial Communications and BusesISA TECHrEXPO Technology Update Conference Proceed-ings v 1 n 3 1997. ISA, Research Triangle Park, NC, USA. p143-150 CODEN: 002744 ISSN: 1054-0032Publication Year: 1997Abstract: If ‘information is the currency of the informationage’, then your Fieldbus field devices like transmitters andpositioners are your printing presses. Fieldbus based controlsystems communicate not only process variables, but also lotsof non-control information. Detailed device self-diagnosticsand automatic failure alarms enables quick determination if aprocess problem is due to the sensor or not, without having todo several field visits. The time that can be saved by nothaving to bring only a few transmitters in to the workshop isenormous. Wear and tear data, exceeded operating conditionsor other abnormal behaviour information may be used forpredictive maintenance scheduling minimizing costly unsched-uled downtime. Field instrument management software willcomplement the basic process control function and reducespaper work and enables automatic record keeping and docu-mentation of configuration, calibration and maintenance. Thisis an important feature to help comply with ISO9000 andISO14000 and other quality and environmental managementsystem requirements for updated and traceable records. Field-bus helps companies increase productivity, flexibility, qualityand comply with ever more stringent environmental regula-tions, and at the same time lower their operating cost. Field-bus is the means whereby companies will stay competitive well

Ž .into the 21 st century. Author abstract 6 Refs. In English EIOrder Number: EIP98013999909Keywords: Man machine systems; Computer control; Com-puter software; Calibration; Quality control; Laws and legisla-tion

Title: ROADMAP TO IMPROVING SOFTWARE PRODUC-TIVITY AND QUALITY

( )Author s : Rozum, James A.Corporate Source: Carnegie Mellon Univ, Pittsburgh, PA,USA