FULSOME: a fuzzy logic modeling tool for software metricians
10 Jun 1999-pp 263-267
TL;DR: While there are many tools available for developing fuzzy models, it is suggested that before there will be real adoption of such techniques by project managers there will need to be suitable tools that support their particular workflows and that use appropriate terminology.
Abstract: There has been a growing body of literature suggesting that some of the problems faced by software development project managers can be at least partially overcome by using fuzzy logic techniques. However, one issue that has been generally overlooked in this recommendation is the means by which these "software metricians" can collect data for, develop, and interpret fuzzy logic models in practice. We describe a freely available system that has been built with this in mind called FULSOME (FUzzy Logic for SOftware MEtrics). While there are many tools available for developing fuzzy models, it is suggested that before there will be real adoption of such techniques by project managers there will need to be suitable tools that support their particular workflows and that use appropriate terminology. Another requirement will be the development of some standard procedures and definitions for such models. Issues involved with membership function elicitation and extraction are also discussed as a first step towards this second goal.
Citations
More filters
08 Mar 2014
TL;DR: The fuzzy logic model fuzzifies the two parts of the COCOMO model i.e. nominal effort prediction and the effort adjustment factor and shows that the performance of the FIS enhanced by increasing the number of membership functions.
Abstract: Software cost estimation is the process of predicting effort required to develop a software system. This effort may be in terms of number of hours of work or number of workers. Precise effort estimation with a high grade of reliability is an indispensable part of effectively software management. Software project costs include the cost incurred in all the expenses, i.e. the cost of project from initiation, development to test, software management, quality management and contingent rework, etc. The imprecision inculcated from the inputs utilized in algorithmic models like constructive cost model COCOMO results in imprecise outputs which leads to erroneous effort estimation. In this paper, a software cost estimation model has been proposed based on fuzzy logic. The fuzzy logic model fuzzifies the two parts of the COCOMO model i.e. nominal effort prediction and the effort adjustment factor. The analysis shows that the performance of the FIS enhanced by increasing the number of membership functions. Validation experiment was carried out on NASA 93 and COCOMO81 public database.
4 citations
Cites methods from "FULSOME: a fuzzy logic modeling too..."
...The predictive capabilities of the FIS were tested using different number of fuzzy sets (11, 13) for input variable size with Gaussian Membership Functions....
[...]
01 Jan 2016
TL;DR: Business processes that involve storing or transmitting personal data are subject to strict regulatory and compliance requirements and the choice of deploying such processes on a shared platform like the cloud hinges on the process owner being convinced that the cloud platform is fully compliant with regulations.
Abstract: Business processes that involve storing or transmitting personal data are subject to strict regulatory and compliance requirements. The choice of deploying such processes on a shared platform like the cloud hinges on the process owner being convinced that the cloud platform is fully compliant with regulations.
3 citations
TL;DR: There is indeed support in the software engineering practitioner community for the use of methods based on the principles of fuzzy logic modeling, particularly if fuzzy logic models are used in a complementary manner with other algorithmic approaches, thus providing a range of predictions as opposed to a single point value.
Abstract: There is a growing body of evidence to suggest that significant benefits may be gained from augmenting current approaches to software development effort estimation, and indeed other project management activities, with models developed using fuzzy logic and other soft computing methods. The tasks undertaken by project managers early in a development process would appear to be particularly amenable to such a strategy, particularly if fuzzy logic models are used in a complementary manner with other algorithmic approaches, thus providing a range of predictions as opposed to a single point value. As well as providing a more intuitively acceptable set of estimates, this would help to reduce or remove the unwarranted level of certainty associated with a point estimate. Furthermore, such an approach would enable organizations to ‘store’ their project management knowledge, making them less susceptible to employee resignations and the like. If fuzzy logic modeling is to be implemented in industry, however, managers must first believe it to be a realistic and workable option. This issue is addressed here by considering two related questions: one, what expectations do project managers have in relation to effort estimation? and two, what is their opinion of the methods that might be useful in this regard? This is followed by a discussion of the results of two surveys of project managers aimed at deriving membership functions using polling methods, the first using an interval declaration approach and the second using votes on fixed points. It is concluded that there is indeed support in the software engineering practitioner community for the use of methods based on the principles of fuzzy logic modeling.
2 citations
Cites background from "FULSOME: a fuzzy logic modeling too..."
...That is to say, most of the work on these modeling methods, including studies undertaken by the authors [13,14,33,34], has been carried out by researchers rather than practitioners, much of it without any definitive sense of the likely acceptability of such methods in an industrial software development setting [11]....
[...]
01 Jan 1999
TL;DR: A clustering based algorithm for automatically deriving fuzzy systems from data (either membership functions, rules, or both may be extracted) is outlined, which assists with the initial system creation process which can be an especially difficult activity for novices.
Abstract: Using fuzzy logic, and associated techniques such as fuzzy clustering, in a teaching environment necessitates the availability of introductory and pedagogically appropriate tools. In a similar manner, introductory level tools may be necessary for practical applications where users are non-specialists in fuzzy theory, as is often the case. For these two scenarios, and many others, the tools that support the use of the modeling technique must satisfy sets of requirements concerning the interface, functionality, and documentation. Examples of these requirements can include the program’s ability to guide the user, without undue restrictions, through the necessary modeling procedures (as in a wizard interface) and explaining (both textually and graphically) the operation of the inference process. After outlining some of the desirable attributes for such tools that are intended to be used by these fuzzy novices, this paper describes the collection of tools collectively known as FUZZYMANAGER. Despite the name, which reflects its project management origins, the system is targeted towards any group of users without a particularly comprehensive or deep knowledge of fuzzy logic, who want an intuitive and graphical approach to fuzzy logic model building. As well as implementing standard inference options and methods, a clustering based algorithm for automatically deriving fuzzy systems from data (either membership functions, rules, or both may be extracted) is outlined. This component assists with the initial system creation process which can be an especially difficult activity for novices.
Cites methods from "FULSOME: a fuzzy logic modeling too..."
...As can be seen, the FUZZYMANAGER software [3] consists of two main application modules, although several others are planned to be added in the next two years....
[...]
01 Jan 1999
TL;DR: The results from two surveys of project managers in New Zealand that asked them various questions about their use of and attitudes towards modelling techniques for supporting the management of software development projects, especially fuzzy logic, are described in this article.
Abstract: This paper describes the results from two surveys of project managers in New Zealand that asked them various questions about their use of and attitudes towards modelling techniques for supporting the management of software development projects, especially fuzzy logic. Each survey is summarized separately and then some overall conclusions are drawn. The results give some indication of how new modelling techniques, and especially fuzzy logic, can be presented to managers. The positive attitude of many managers towards the use of fuzzy logic can be used within their current software development management practices.
Cites background from "FULSOME: a fuzzy logic modeling too..."
...Some of these organizations are now being approached to evaluate the FULSOME system, as described in [3], in a more practical setting....
[...]
References
More filters
Book•
01 Jan 2013
TL;DR: The Second Edition of Software Metrics provides an up-to-date, coherent, and rigorous framework for controlling, managing, and predicting software development processes.
Abstract: From the Publisher:
The Second Edition of Software Metrics provides an up-to-date, coherent, and rigorous framework for controlling, managing, and predicting software development processes. With an emphasis on real-world applications, Fenton and Pfleeger apply basic ideas in measurement theory to quantify software development resources, processes, and products. The book offers an accessible and comprehensive introduction to software metrics, now an essential component of software engineering for both classroom and industry. Software Metrics features extensive case studies from Hewlett Packard, IBM, the U.S. Department of Defense, Motorola, and others, in addition to worked examples and exercises. The Second Edition includes up-to-date material on process maturity and measurement, goal-question-metric, planning a metrics program, measurement in practice, experimentation, empirical studies, ISO9216, and metric tools.
2,827 citations
"FULSOME: a fuzzy logic modeling too..." refers background in this paper
...Software metrics is the field of research and practice that involves investigating the characteristics of and relationships between sets of attributes associated with software development projects, usually in terms of products, processes, and resources [1]....
[...]
TL;DR: The use of regression analysis to derive predictive equations for software metrics has recently been complemented by increasing numbers of studies using non-traditional methods, such as neural networks, fuzzy logic models, case-based reasoning systems, and regression trees.
Abstract: The use of regression analysis to derive predictive equations for software metrics has recently been complemented by increasing numbers of studies using non-traditional methods, such as neural networks, fuzzy logic models, case-based reasoning systems, and regression trees. There has also been an increasing level of sophistication in the regression-based techniques used, including robust regression methods, factor analysis, and more effective validation procedures. This paper examines the implications of using these methods and provides some recommendations as to when they may be appropriate. A comparison of the various techniques is also made in terms of their modelling capabilities with specific reference to software metrics.
201 citations
TL;DR: A new consistent and robust method called the least-squares of inverted balanced relative errors (LIRS) is proposed and its superiority to the ordinary least-Squares method is demonstrated by use of five actual data sets.
178 citations
"FULSOME: a fuzzy logic modeling too..." refers background in this paper
...Even once sufficient quantities of data are available, data purity is generally difficult to ascertain – necessitating some treatment of unusual observations [4]....
[...]
21 Sep 1997
TL;DR: Software metrics are measurements of the software development process and product that can be used as variables (both dependent and independent) in models for project management and the use of alternative techniques, especially fuzzy logic, is investigated and some usage recommendations are made.
Abstract: Software metrics are measurements of the software development process and product that can be used as variables (both dependent and independent) in models for project management. The most common types of these models are those used for predicting the development effort for a software system based on size, complexity, developer characteristics, and other metrics. Despite the financial benefits from developing accurate and usable models, there are a number of problems that have not been overcome using the traditional techniques of formal and linear regression models. These include the nonlinearities and interactions inherent in complex real-world development processes, the lack of stationarity in such processes, over-commitment to precisely specified values, the small quantities of data often available, and the inability to use whatever knowledge is available where exact numerical values are unknown. The use of alternative techniques, especially fuzzy logic, is investigated and some usage recommendations are made.
87 citations
"FULSOME: a fuzzy logic modeling too..." refers background in this paper
...Fuzzy logic has recently gained a greater amount of attention in the software metrics literature as a means for solving some of these long-standing problems [2, 3]....
[...]
Journal Article•
24 citations