Narrow Results

High software reliability is an important attribute of high-assurance systems. Software quality models yield timely predictions of quality indicators on a module-by-module basis, enabling one to focus on finding faults early in development. This paper introduces the Classification And Regression Trees (CART) a algorithm to practitioners in high-assurance systems engineering. This paper presents practical lessons learned on building classification trees for software quality modeling, including an innovative way to control the balance between misclassification rates. A case study of a very large telecommunications system used CART to build software quality models. The models predicted whether or not modules would have faults discovered by customers, based on various sets of software product and process metrics as independent variables. We found that a model based on two software product metrics had comparable accuracy to a model based on forty product and process metrics.

"Knowledge discovery in data bases" (KDD) for software engineering is a process for finding useful information in the large volumes of data that are a byproduct of software development, such as data bases for configuration management and for problem reporting. This paper presents guidelines for extracting innovative process metrics from these commonly available data bases. This paper also adapts the Classification And Regression Trees algorithm, CART, to the KDD process for software engineering data. To our knowledge, this algorithm has not been used previously for empirical software quality modeling. In particular, we present an innovative way to control the balance between misclassification rates. A KDD case study of a very large legacy telecommunications software system found that variables derived from source code, configuration management transactions, and problem reporting transactions can be useful predictors of software quality. The KDD process discovered that for this software development environment, out of forty software attributes, only a few of the predictor variables were significant. This resulted in a model that predicts whether modules are likely to have faults discovered by customers. Software developers need such predictions early in development to target software enhancement techniques to the modules that need improvement the most.

Embedded-computer systems have become essential to life in modern society. For example, the backbone of society's information infrastructure is telecommunications. Embedded systems must have highly reliable software, so that we avoid the severe consequences of failures, intolerable down-time, and expensive repairs in remote locations. Moreover, today's fast-moving technology marketplace mandates that embedded systems evolve, resulting in multiple software releases embedded in multiple products.

Software quality models can be valuable tools for software engineering of embedded systems, because some software-enhancement techniques are so expensive or time-consuming that it is not practical to apply them to all modules. Targeting such enhancement techniques is an effective way to reduce the likelihood of faults discovered in the field. Research has shown software metrics to be useful predictors of software faults. A software quality model is developed using measurements and fault data from a past release. The calibrated model is then applied to modules currently under development. Such models yield predictions on a module-by-module basis.

This paper examines the Classification And Regression Trees (CART) algorithm for building tree-based models that predict which software modules have high risk of faults to be discovered during operations. CART is attractive because it emphasizes pruning to achieve robust models. This paper presents details on the CART algorithm in the context of software engineering of embedded systems. We illustrate this approach with a case study of four consecutive releases of software embedded in a large telecommunications system. The level of accuracy achieved in the case study would be useful to developers of an embedded system. The case study indicated that this model would continue to be useful over several releases as the system evolves.

Reliable software is mandatory for complex mission-critical systems. Classifying modules as fault-prone, or not, is a valuable technique for guiding development processes, so that resources can be focused on those parts of a system that are most likely to have faults.

Logistic regression offers advantages over other classification modeling techniques, such as interpretable coefficients. There are few prior applications of logistic regression to software quality models in the literature, and none that we know of account for prior probabilities and costs of misclassification. A contribution of this paper is the application of prior probabilities and costs of misclassification to a logistic regression-based classification rule for a software quality model.

This paper also contributes an integrated method for using logistic regression in software quality modeling, including examples of how to interpret coefficients, how to use prior probabilities, and how to use costs of misclassifications. A case study of a major subsystem of a military, real-time system illustrates the techniques.

Because highly reliable software is becoming an essential ingredient in many systems, software developers apply various techniques to discover faults early in development, such as more rigorous reviews, more extensive testing, and strategic assignment of key personnel. Our goal is to target reliability enhancement activities to those modules that are most likely to have problems. This paper presents a methodology that incorporates genetic programming for predicting the order of software modules based on the expected number of faults. This is the first application of genetic programming to software engineering that we know of. We found that genetic programming can be used to generate software quality models whose inputs are software metrics collected earlier in development, and whose output is a prediction of the number of faults that will be discovered later in development or during operations. We established ordinal evaluation criteria for models, and conducted an industrial case study of software from a military communications system. Case study results were sufficiently good to be useful to a project for choosing modules for extra reliability enhancement treatment.