Accurate project effort prediction is an important goal for the software engineering community. To date most work has focused upon building algorithmic models of effort, example COCOMO. These can be calibrated to local environments. We describe alternative approach estimation based use analogies. The underlying principle characterize projects in terms features (for example, number interfaces, development method or size functional requirements document). Completed are stored and then problem...

Background--Self-evidently empirical analyses rely upon the quality of their data. Likewise, replications accurate reporting and using same rather than similar versions datasets. In recent years, there has been much interest in machine learners to classify software modules into defect-prone not categories. The publicly available NASA datasets have extensively used as part this research. Objective--This short note investigates extent which published based on defect are meaningful comparable....

The paper aims to provide the software estimation research community with a better understanding of meaning of, and relationship between, two statistics that are often used assess accuracy predictive models: mean magnitude relative error, MMRE, number predictions within 25% actuals, pred(25). It is demonstrated MMRE pred(25) are, respectively, measures spread kurtosis variable z where z=estimate/actual. Thus, considered be measure accuracy, such as properties distribution z. suggested...

BACKGROUND - Predicting defect-prone software components is an economically important activity and so has received a good deal of attention. However, making sense the many, sometimes seemingly inconsistent, results difficult. OBJECTIVE We propose evaluate general framework for defect prediction that supports 1) unbiased 2) comprehensive comparison between competing systems. METHOD The comprised scheme evaluation components. analyzes performance learning schemes given historical data sets....

Background. The ability to predict defect-prone software components would be valuable. Consequently, there have been many empirical studies evaluate the performance of different techniques endeavouring accomplish this effectively. However no one technique dominates and so designing a reliable defect prediction model remains problematic. Objective. We seek make sense conflicting experimental results understand which factors largest effect on predictive performance. Method. conduct...

Context: Software defect prediction (SDP) is an important challenge in the field of software engineering, hence much research work has been conducted, most notably through use machine learning algorithms. However, class-imbalance typified by few defective components and many non-defective ones a common occurrence causing difficulties for these methods. Imbalanced aims to deal with this problem recently deployed some researchers, unfortunately inconsistent results. Objective: We conduct...

Metaheuristic techniques such as genetic algorithms, simulated annealing and tabu search have found wide application in most areas of engineering. These also been applied business, financial economic modelling. Metaheuristics to three software engineering: test data generation, module clustering cost/effort prediction, yet there remain many engineering problems which be tackled using metaheuristics. It is surprising that metaheuristics not more widely engineering; are characterised by...

The need for accurate software prediction systems increases as becomes much larger and more complex. We believe that the underlying characteristics: size, number of features, type distribution, etc., data set influence choice system to be used. For this reason, we would like control characteristics such sets in order systematically explore relationship between accuracy, system, characteristic. It also useful have a large validation set. Our solution is simulate allowing both possibility...

The staff resources or effort required for a software project are notoriously difficult to estimate in advance. To date most work has focused upon algorithmic cost models such as COCOMO and Function Points. These can suffer from the disadvantage of need calibrate model each individual measurement environment coupled with very variable accuracy levels even after calibration. An alternative approach is use analogy estimation. We demonstrate that this method considerable promise we show it out...

The paper describes an empirical investigation into industrial object oriented (OO) system comprised of 133000 lines C++. was a subsystem telecommunications product and developed using the Shlaer-Mellor method (S. Shlaer S.J. Mellor, 1988; 1992). From this study, we found that there little use OO constructs such as inheritance, therefore polymorphism. It also significant difference in defect densities between those classes participated inheritance structures did not, with former being...

Empirical studies on software prediction models do not converge with respect to the question "which model is best?" The reason for this lack of convergence poorly understood. In simulation study, we have examined a frequently used research procedure comprising three main ingredients: single data sample, an accuracy indicator, and cross validation. Typically, these empirical compare machine learning regression model. our use results suggest that it itself unreliable. This reliability may...

McCabe's cyclomatic complexity metric is widely cited as a useful predictor of various software attributes such reliability and development effort. This critique demonstrates that it based upon poor theoretical foundations an inadequate model development. The argument the provides developer with engineering approximation not borne out by empirical evidence. Furthermore, would appear for large class no more than proxy for, in many cases outperformed by, lines code.

Much current software defect prediction work focuses on the number of defects remaining in a system. In this paper, we present association rule mining based methods to predict associations and correction effort. This is help developers detect assist project managers allocating testing resources more effectively. We applied proposed SEL data consisting than 200 projects over 15 years. The results show that, for prediction, accuracy very high false-negative rate low. Likewise, effort both...

Context: There is considerable diversity in the range and design of computational experiments to assess classifiers for software defect prediction. This particularly so, regarding choice classifier performance metrics. Unfortunately some widely used metrics are known be biased, particular F1.

The staff resources or effort required for a software project are notoriously difficult to estimate in advance. To date most work has focused upon algorithmic cost models such as COCOMO and Function Points. These can suffer from the disadvantage of need calibrate model each individual measurement environment coupled with very variable accuracy levels even after calibration. An alternative approach is use analogy estimation. We demonstrate that this method considerable promise we show it out...

A suite of object oriented software metrics has recently been proposed by S.R. Chidamber and C.F. Kemerer (see ibid., vol. 20, p. 476-94, 1994). While the authors have taken care to ensure their a sound measurement theoretical basis, we argue that is premature begin applying such while there remains uncertainty about precise definitions many quantities be observed impact upon subsequent indirect metrics. In particular, show some ambiguities associated with seemingly simple concept number...

OBJECTIVE - to assess the extent and types of techniques used manage quality within software engineering data sets. We consider this a particularly interesting question in context initiatives promote sharing secondary analysis METHOD we perform systematic review available empirical studies. RESULTS only 23 out many hundreds studies assessed, explicitly considered quality. CONCLUSIONS first, community needs appropriateness set being utilised; not all sets are equal. Second, need more research...

BACKGROUND-The systematic review is becoming a more commonly employed research instrument in empirical software engineering. Before undue reliance placed on the outcomes of such reviews it would seem useful to consider robustness approach this particular context. OBJECTIVE-The aim study assess reliability as instrument. In particular, we wish investigate consistency process and stability outcomes. METHOD-We compare results two independent undertaken with common question. RESULTS-The find...