Software measurement has the potential to play an important role in risk management during product development. Metrics incorporated into predictive models can give advanced warning of potential risks. However, the common approach of using simple regression models, notably to predict software defects, can lead to inappropriate risk management decisions. These nave models should be replaced with predictive models incorporating genuine cause-effect relationships. We show how these can be built using Bayesian networks; a powerful graphical modelling technique. We describe how a Bayesian network for software quality risk management is providing accurate predictions of software defects in a range of real projects. As well as their use for prediction, Bayesian networks can also be used for performing a range of "what if" scenarios to identify potential problems and possible improvement actions. This really is the dawn of an exciting new era for software measurement.

© 2004 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in

Abstract — In today's embedded software systems development, non-functional requirements (e.g., dependability, maintainability) are becoming more and more important. Simultaneously the increasing pressure to develop software in less time and at lower costs pushes software industry towards product line’s solutions. To support product lines for high quality embedded software, quality models are needed. In this paper, we investigate to which extent existing quality modeling approaches facilitate high quality software product lines. First, we define several requirements for an appropriate quality model. Then, we use those requirements to review the existing quality modeling approaches. We conclude from the review that no single quality model fulfills all of our requirements. However, several approaches contain valuable characteristics. Based upon those characteristics, we propose the Prometheus approach. Prometheus is a goal-oriented method that integrates quantitative and qualitative approaches to quality control. The method starts quality modeling early in the software lifecycle and is reusable across product lines.

This paper shows how recent revolutionary Bayesian Network (BN) algorithms can be used to model very complex reliability problems in a simple unified way. The algorithms work for socalled hybrid BNs, which are BNs that can contain a mixture of both discrete and continuous variables. Such hybrid BNs enable us to model failure times and reliability together. The approach allows a compact representation of the event-dependent failure behaviours characteristic of fault-tolerant systems, avoiding the state space explosion problem of the Markov Chain based approaches. The BN framework presented is able to solve any configuration of static and dynamic gates with general time-to-failure distributions, without using numerical integration techniques or simulation methods. Unlike other approaches (which tend to be restricted to using exponential distributions) we can use as input any parametric or empirical failure rate distribution. The approach offers a powerful framework for analysts and decision makers to successfully perform robust reliability assessment. Sensitivity, uncertainty, diagnosis analysis, common cause failures, and warranty analysis can also be easily performed within this framework.

A risk-based decision-making methodology conceived and developed at JPL and NASA has been used to aid in decision making for spacecraft technology assessment, adoption, development and operation. It takes a riskcentric perspective, through which risks are used as a reasoning step to interpose between mission objectives and risk mitigation measures. The novel aspects of this methodology lie in: Broad-ranging treatment of objectives, risks and mitigations: objectives encompass science objectives of the mission, development considerations, and constraints

There has been extensive research in the area of data mining over the last decade, but relatively little research in algorithmic mining. Some researchers shun the idea of incorporating explicit knowledge with a Genetic Program environment. At best, very domain specific knowledge is hard wired into the GP modeling process. This work proposes a new approach called the Genetically Engineerable Evolvable Program (GEEP). In this approach, explicit knowledge is made available to the GP. It is considered breadth-based, in that all pieces of knowledge are independent of each other. Several experiments are performed on a NASA-based data set using established equations from other researchers in order to predict software defects. All results are statistically validated.

Intelligent software estimation models are need of the time. With increased development of Bayesian networks for software project management, one requires an explicit Bayesian Network (BN) to provide effort estimates based on historical data. This paper proposes a simple BN, based on classification approach. However the classes of ranges of size value, are distributed with help of fuzzification to distribute the probability of crisp value The model is simple and smaller, thus can easily be connected to static as well as dynamic Bayesian Networks. General Terms Software effort estimation.