Software Project Dataset Research Articles

Predicting the Number of Software Faults using Deep Learning

The software testing phase requires considerable time, effort, and cost, particularly when there are many faults. Thus, developers focus on the evolution of Software Fault Prediction (SFP) to predict faulty units in advance, therefore, improving software quality significantly. Forecasting the number of faults in software units can efficiently direct software testing efforts. Previous studies have employed several machine learning models to determine whether a software unit is faulty. In this study, a new, simple deep neural network approach that can adapt to the type of input data was designed, utilizing Convolutional Neural Networks (CNNs) and Multi-Layer Perceptron (MLP), to predict the number of software faults. Twelve open-source software project datasets from the PROMISE repository were used for testing and validation. As data imbalance can negatively impact prediction accuracy, the new version of synthetic minority over-sampling technique (SMOTEND) was used to resolve data imbalance. In experimental results, a lower error rate was obtained for MLP, compared to CNN, reaching 0.195, indicating the accuracy of this prediction model. The proposed approach proved to be effective when compared with two of the best machine learning models in the field of prediction. The code will be available on GitHub.

An efficient heuristic algorithm for software module clustering optimization

In the lifecycle of software projects, maintenance tasks usually entail 75% of the total costs, where most efforts are spent in understanding the program. To improve the maintainability of software projects, the code is often divided into components, which are then grouped in different modules following good design principles, lowering coupling and increasing cohesion. The Software Module Clustering Problem (SMCP) is an optimization problem that looks for maximizing the modularity of software projects in the context of Search-Based Software Engineering. In the SMCP, projects are often modeled as graphs. Therefore, the SMCP can be interpreted as a graph partitioning problem, which is proved to be NP-hard. In this work, we propose a new heuristic algorithm for software modularization, based on a Greedy Randomized Adaptive Search Procedure with Variable Neighborhood Descent. We present a three-fold categorization of neighborhoods for the SMCP and leverage domain-specific information to filter unpromising solutions. Our proposal has been successfully tested over a dataset of real software projects, outperforming the previous state-of-the-art approach in terms of Modularization Quality in very short computing times. Therefore, it could be integrated in software development tools to improve the quality of software projects in real time.

Hybrid PSO-SA approach for feature weighting in analogy-based software project effort estimation

Software effort estimation plays an important role in software project management, and analogy-based estimation (ABE) is the most common method used for this purpose. ABE estimates the effort required for a new software project based on its similarity to previous projects. A similarity between the projects is evaluated based on a set of project features, each of which has a particular effect on the degree of similarity between projects and the effort feature. The present study examines the hybrid PSO-SA approach for feature weighting in analogy-based software project effort estimation. The proposed approach was implemented and tested on two well-known datasets of software projects. The performance of the proposed model was compared with other optimization algorithms based on MMRE, MDMRE, and PRED(0.25) measures. The results showed that weighted ABE models provide more accurate and better effort estimates relative to unweighted ABE models and that the PSO-SA hybrid approach has led to better and more accurate results compared with the other weighting approaches in both datasets.

Effort prediction for the software project construction phase

AbstractThe construction phase effort prediction is needed for assigning resources to teams of practitioners destined specifically to this phase of the software development life cycle (SDLC). Construction effort (CE) has been reported between 27.5% and 58% of the total SDLC effort causing the uncertainty of taking these percentages as reference. A support vector regression (SVR) training involves quadratic programming problems that can analytically be solved using a sequential minimal optimization (SMO) algorithm. Moreover, a Pearson VII (PUK) kernel is useful to replace a set of kernel functions commonly used by a SVR. The objective of this study is to apply the SMO with the PUK to train SVR for predicting CE. The SVR model trained with the SMO algorithm having as kernel to the PUK (SVR‐SMO‐PUK) prediction accuracy was statistically compared to those accuracies obtained from statistical regression (SR), neural network (NN), and two types of SVR. Seven international public data sets of software projects were used. Results showed that the SVR‐SMO‐PUK was better than the SR in five data sets and better than NN in two of these five data sets. It was equal than SR and NN in the remaining two data sets. It was equal than ε‐SVR and ʋ‐SVR in the seven data sets. Thus, the SVR‐SMO‐PUK is useful to software managers to predict CE.

Software defect prediction with imbalanced distribution by radius‐synthetic minority over‐sampling technique

AbstractSoftware defect prediction, which can identify the defect‐prone modules, is an effective technology to ensure the quality of software products. Due to the importance in software maintenance, many learning‐based software defect prediction models are presented in recent years. Actually, the defects usually occupy a very small proportions in software source codes; thus, the imbalanced distributions between defect‐prone modules and non‐defect‐prone modules increase the learning difficulty of the classification task. To address this issue, we present a random over‐sampling mechanism used to generate minority‐class samples from high‐dimensional sampling space to deal with the imbalanced distributions in software defect prediction, in which two constraints are applied to provide a robust way to generate new synthetic samples, that is, scaling the random over‐sampling scope to a reasonable area and distinguishing the majority‐class samples in a critical region. Based on nine open datasets of software projects, we experimentally verify that our presented method is effective on predict the defect‐prone modules, and the effect is superior to the traditional imbalanced processing methods.

Automatic team recommendation for collaborative software development

In large-scale collaborative software development, building a team of software practitioners can be challenging, mainly due to overloading choices of candidate members to fill in each role. Furthermore, having to understand all members’ diverse backgrounds, and anticipate team compatibility could significantly complicate and attenuate such a team formation process. Current solutions that aim to automatically suggest software practitioners for a task merely target particular roles, such as developers, reviewers, and integrators. While these existing approaches could alleviate issues presented by choice overloading, they fail to address team compatibility while members collaborate. In this paper, we propose RECAST, an intelligent recommendation system that suggests team configurations that satisfy not only the role requirements, but also the necessary technical skills and teamwork compatibility, given task description and a task assignee. Specifically, RECAST uses Max-Logit to intelligently enumerate and rank teams based on the team-fitness scores. Machine learning algorithms are adapted to generate a scoring function that learns from heterogenous features characterizing effective software teams in large-scale collaborative software development. RECAST is evaluated against a state-of-the-art team recommendation algorithm using three well-known open-source software project datasets. The evaluation results are promising, illustrating that our proposed method outperforms the baselines in terms of team recommendation with 646% improvement (MRR) using the exact-match evaluation protocol.

Optimized COCOMO parameters using hybrid particle swarm optimization

Software effort and cost estimation are crucial parts of software project development. It determines the budget, time, and resources needed to develop a software project. The success of a software project development depends mainly on the accuracy of software effort and cost estimation. A poor estimation will impact the result, which worsens the project management. Various software effort estimation model has been introduced to resolve this problem. COnstructive COst MOdel (COCOMO) is a well-established software project estimation model; however, it lacks accuracy in effort and cost estimation, especially for current projects. Inaccuracy and complexity in the estimated effort have made it difficult to efficiently and effectively develop software, affecting the schedule, cost, and uncertain estimation directly. In this paper, Particle Swarm Optimization (PSO) is proposed as a metaheuristics optimization method to hybrid with three traditional state-of-art techniques such as Support Vector Machine (SVM), Linear Regression (LR), and Random Forest (RF) for optimizing the parameters of COCOMO models. The proposed approach is applied to the NASA software project dataset downloaded from the promise repository. Comparing the proposed approach has been made with the three traditional algorithms; however, the obtained results confirm low accuracy before hybrid with PSO. Overall, the results showed that PSOSVM on the NASA software project dataset could improve effort estimation accuracy and outperform other models.

Particle Swarm Optimization for Predicting the Development Effort of Software Projects

Software project planning includes as one of its main activities software development effort prediction (SDEP). Effort (measured in person-hours) is useful to budget and bidding the projects. It corresponds to one of the variables most predicted, actually, hundreds of studies on SDEP have been published. Therefore, we propose the application of the Particle Swarm Optimization (PSO) metaheuristic for optimizing the parameters of statistical regression equations (SRE) applied to SDEP. Our proposal incorporates two elements in PSO: the selection of the SDEP model, and the automatic adjustment of its parameters. The prediction accuracy of the SRE optimized through PSO (PSO-SRE) was compared to that of a SRE model. These models were trained and tested using eight data sets of new and enhancement software projects obtained from an international public repository of projects. Results based on statistically significance showed that the PSO-SRE was better than the SRE in six data sets at 99% of confidence, in one data set at 95%, and statistically equal than SRE in the remaining data set. We can conclude that the PSO can be used for optimizing SDEP equations taking into account the type of development, development platform, and programming language type of the projects.

Empirical Evaluation of Mimic Software Project Data Sets for Software Effort Estimation

To conduct empirical research on industry software development, it is necessary to obtain data of real software projects from industry. However, only few such industry data sets are publicly available; and unfortunately, most of them are very old. In addition, most of today's software companies cannot make their data open, because software development involves many stakeholders, and thus, its data confidentiality must be strongly preserved. To that end, this study proposes a method for artificially generating a “mimic” software project data set, whose characteristics (such as average, standard deviation and correlation coefficients) are very similar to a given confidential data set. Instead of using the original (confidential) data set, researchers are expected to use the mimic data set to produce similar results as the original data set. The proposed method uses the Box-Muller transform for generating normally distributed random numbers; and exponential transformation and number reordering for data mimicry. To evaluate the efficacy of the proposed method, effort estimation is considered as potential application domain for employing mimic data. Estimation models are built from 8 reference data sets and their concerning mimic data. Our experiments confirmed that models built from mimic data sets show similar effort estimation performance as the models built from original data sets, which indicate the capability of the proposed method in generating representative samples.

Ensemble learning for software fault prediction problem with imbalanced data

Fault prediction problem has a crucial role in the software development process because it contributes to reducing defects and assisting the testing process towards fault-free software components. <span lang="EN-US">Therefore, there are a lot of efforts aiming to address this type of issues, in which static code characteristics are usually adopted to construct fault classification models. </span> One of the challenging problems influencing the performance of predictive classifiers is the high imbalance among patterns belonging to different classes. This paper aims to integrate the sampling techniques and common classification techniques to form a useful ensemble model for the software defect prediction problem. The empirical results conducted on the benchmark datasets of software projects have shown the promising performance of our proposal in comparison with individual classifiers.

Uncertainty management in software effort estimation using a consistent fuzzy analogy-based method

Software effort estimation is a critical task in software project development management. Unfortunately, the uncertainty and inaccuracy are inherent properties of the software effort estimation environment. These are caused by the limited capabilities of the managers, to foresee, measure and describe factors influencing the software effort. The promising Fuzzy Analogy-based Software Effort Estimation model (FASEE) employs successfully fuzzy logic with approximate reasoning theory to handle imprecision and reasoning under uncertainty. Also, FASEE use possibility distribution to quantify the uncertainty in the estimate that aid the software managers to assess risks. Yet, the FASEE suffer from the low data quality and the uncertainty induced in the reasoning process. In this paper, we propose an enhancement of the FASEE, by imposing consistency criteria to deal with the aforementioned drawbacks. So, the underlying model, called Consistent Fuzzy Analogy-based Software Effort Estimation (C-FASEE) is endowed with two capabilities. The first one introduces consistency criteria in attribute representation by fuzzy sets to enable fitting each attribute to the software effort. The second one introduces a new relation of confidence to measure the extent that the resulted most similar projects respect the assumption “similar projects have similar efforts”. Moreover, the C-FASEE method provide a fuzzy estimate of the most possible fuzzy set will the true effort of the new software project falls in. This allow to the software manager to assess risks more optimally. The proposed C-FASEE is validated over thirteen software project datasets that represent different complexities. The obtained results are compared to variant methods of the analogy-based software effort estimation approach. The experimental results show that our proposal provides a good estimation accuracy of and has significantly best performance against the comparison methods.

Research and Appalication of Software Defect Predictionn based on BP-Migration learning

Software Defect Prediction has been an important part of Software engineering research since the 1970s. This technique is used to calculate and analyze the measurement and defect information of the historical software module to complete the defect prediction of the new software module. Currently, most software defect prediction model is established on the basis of the same software project data set. The training date sets used to construct the model and the test data sets used to validate the model are from the same software projects. But in practice, for those has less historical data of a software project or new projects, the defect of traditional prediction method shows lower forecast performance. For the traditional method, when the historical data is insufficient, the software defect prediction model cannot be fully studied. It is difficult to achieve high prediction accuracy. In the process of cross-project prediction, the problem that we will faced is data distribution differences. For the above problems, this paper presents a software defect prediction model based on migration learning and traditional software defect prediction model. This model uses the existing project data sets to predict software defects across projects. The main work of this article includes: 1) Data preprocessing. This section includes data feature correlation analysis, noise reduction and so on, which effectively avoids the interference of over-fitting problem and noise data on prediction results. 2) Migrate learning. This section analyzes two different but related project data sets and reduces the impact of data distribution differences. 3) Artificial neural networks. According to class imbalance problems of the data set, using artificial neural network and dynamic selection training samples reduce the influence of prediction results because of the positive and negative samples data. The data set of the Relink project and AEEEM is studied to evaluate the performance of the f-measure and the ROC curve and AUC calculation. Experiments show that the model has high predictive performance.

Software Cost Estimation Using Environmental Adaptation Method

Environmental adaptation method (EAM) is one of the evolutionary algorithms for solving single objective optimization problems. After the first proposal of EAM, other variants have been suggested to speed up the convergence and to maintain the population diversity. Among them, IEAM-RP works with real numbers and was able to achieve the desired goal during the optimization process. In this paper, IEAM-RP is used to predict the effort required to develop the software product. The experiments are carried out on NASA software project dataset to check the effectiveness of IEAM-RP. The experimental results demonstrated that the overall performance of IEAM-RP is quite satisfactory in predicting the effort required to develop a software.

Case-based reasoning with optimized weight derived by particle swarm optimization for software effort estimation

Software effort estimation (SEE) is the process of forecasting the effort required to develop a new software system, which is critical to the success of software project management and plays a significant role in software management activities. This study examines the potentials of the SEE method by integrating particle swarm optimization (PSO) with the case-based reasoning (CBR) method, where the PSO method is adopted to optimize the weights in weighted CBR. The experiments are implemented based on two datasets of software projects from the Maxwell and Desharnais datasets. The effectiveness of the proposed model is compared with other published results in terms of the performance measures, which are MMRE, Pred(0.25), and MdMRE. Experimental results show that the weighed CBR generates better software effort estimates than the unweighted CBR methods, and PSO-based weighted grey relational grade CBR achieves better performance and robustness in both datasets than other popular methods.

A multi-release software reliability modeling for open source software incorporating dependent fault detection process

The increasing dependence of our modern society on software systems has driven the development of software products become even more competitive and time-consuming. Single release software product no longer meets the increasing market requirements. Thereby it is important to release multiple version software products in order to add new features in the next release and fix remaining faults from previous release. In this paper, we develop a multi-release software reliability model with consideration of the remaining software faults from previous release and the new introduced-faults (from newly added features). Additionally, dependent fault detection process is taken into account in this research. In particular, the detection of a new fault for developing the next release depends on the detection of the remaining faults from previous release and the detection of the new introduced-faults. The proposed model is validated on the open source software project datasets with multiple releases.

Cross-validation based K nearest neighbor imputation for software quality datasets: An empirical study

Being able to predict software quality is essential, but also it pose significant challenges in software engineering. Historical software project datasets are often being utilized together with various machine learning algorithms for fault-proneness classification. Unfortunately, the missing values in datasets have negative impacts on the estimation accuracy and therefore, could lead to inconsistent results. As a method handling missing data, K nearest neighbor (KNN) imputation gradually gains acceptance in empirical studies by its exemplary performance and simplicity. To date, researchers still call for optimized parameter setting for KNN imputation to further improve its performance. In the work, we develop a novel incomplete-instance based KNN imputation technique, which utilizes a cross-validation scheme to optimize the parameters for each missing value. An experimental assessment is conducted on eight quality datasets under various missingness scenarios. The study also compared the proposed imputation approach with mean imputation and other three KNN imputation approaches. The results show that our proposed approach is superior to others in general. The relatively optimal fixed parameter settings for KNN imputation for software quality data is also determined. It is observed that the classification accuracy is improved or at least maintained by using our approach for missing data imputation.

An Investigation into the Suitability of k-Nearest Neighbour (k-NN) for Software Effort Estimation

Software effort estimation is an increasingly significant field, due to the overwhelming role of software in today’s global market. Effort estimation involves forecasting the effort in person-months or hours required for developing a software. It is vital to ideal planning and paramount for controlling the software development process. However, there is presently no optimal method to accurately estimate the effort required to develop a software system. Inaccurate estimation leads to poor use of resources and perhaps failure of the software project. Effort estimation also plays a key role in deducing cost of a software project. Software cost estimation includes the generation of the effort estimates and project duration to predict cost required to develop software project. Thus, effort is very essential and there is always need to enhance the accuracy as much as possible. This study evaluates and compares the potential of Constructive COst MOdel II (COCOMO II) and k-Nearest Neighbor (k-NN) on software project dataset. By the analysis of results received from each method, it may be concluded that the proposed method k-NN yields better performance over the other technique utilized in this study.

The ISBSG Software Project Repository: An Analysis from Six Sigma Measurement Perspective for Software Defect Estimation

The International Software Benchmarking Standards Group (ISBSG) provides to researchers and practitioners a repository of software projects’ data that has been used to date mostly for benchmarking and project estimation purposes, but rarely for software defects analysis. Sigma, in statistics, measures how far a process deviates from its goal. Six Sigma focuses on reducing variations within processes, because such variations may lead to an inconsistency in achieving projects’ specifications which represent “defects”, which mean not meeting customers’ satisfaction. Six Sigma provides two methodologies to solve organizations’ problems: “Define-Measure-Analyze-Improve-Control” process cycle (DMAIC) and Design of Six Sigma (DFSS). The DMAIC focuses on improving the existed processes, while the DFSS focuses on redesigning the existing processes and developing new processes. This paper presents an approach to provide an analysis of ISBSG repository based on Six Sigma measurements. It investigates the use of the ISBSG data repository with some of the related Six Sigma measurement aspects, including Sigma defect measurement and software defect estimation. This study presents the dataset preparation consisting of two levels of data preparations, and then analyzed the quality-related data fields in the ISBSG MS-Excel data extract (Release 12 - 2013). It also presents an analysis of the extracted dataset of software projects. This study has found that the ISBSG MS-Excel data extract has a high ratio of missing data within the data fields of “Total Number of Defects” variable, which represents a serious challenge when the ISBSG dataset is being used for software defect estimation.

An empirical study of some software fault prediction techniques for the number of faults prediction

During the software development process, prediction of the number of faults in software modules can be more helpful instead of predicting the modules being faulty or non-faulty. Such an approach may help in more focused software testing process and may enhance the reliability of the software system. Most of the earlier works on software fault prediction have used classification techniques for classifying software modules into faulty or non-faulty categories. The techniques such as Poisson regression, negative binomial regression, genetic programming, decision tree regression, and multilayer perceptron can be used for the prediction of the number of faults. In this paper, we present an experimental study to evaluate and compare the capability of six fault prediction techniques such as genetic programming, multilayer perceptron, linear regression, decision tree regression, zero-inflated Poisson regression, and negative binomial regression for the prediction of number of faults. The experimental investigation is carried out for eighteen software project datasets collected from the PROMISE data repository. The results of the investigation are evaluated using average absolute error, average relative error, measure of completeness, and prediction at level l measures. We also perform Kruskal–Wallis test and Dunn’s multiple comparison test to compare the relative performance of the considered fault prediction techniques.

A Novel Technique of Optimization for the COCOMO II Model Parameters using Teaching-Learning-Based Optimization Algorithm

Software cost estimation is a critical activity in the development life cycle for controlling risks and planning project schedules. Accurate estimation of the cost before the start-up of a project is essential for both the developers and the customers. Therefore, many models were proposed to address this issue, in which COCOMO II has been being widely employed in actual software projects. Good estimation models, such as COCOMO II, can avoid insufficient resources being allocated to a project. However, parameters for estimation formula in this model have not been optimized yet, and so the estimated results are not close to the actual results. In this paper, a novel technique to optimize the coefficients for COCOMO II model by using teaching-learning-based optimization (TLBO) algorithm is proposed. The performance of the model after optimizing parameters was tested on NASA software project dataset. The obtained results indicated that the improvement of parameters provided a better estimation capabilities compared to the original COCOMO II model.