Software Maintainability Prediction Models Research Articles

A Hyper-parameter Tuning based Novel Model for Prediction of Software Maintainability

Software maintainability is regarded as one of the most important characteristics of any software system. In today's digital world, the expanding significance of software maintenance is motivating the development of efficient software maintainability prediction (SMP) models using statistical and machine learning methods. This study proposes a hyper-parameter optimizable Software Maintainability Prediction (HPOSMP) model using the hybridized approach of data balancing and hyper-parameter optimization of Machine Learning (ML) approach using software maintainability datasets. The training dataset has been created with object-oriented software namely UIMS and QUES. To balance the dataset, Synthetic Minority Oversampling Technique (SMOTE) technology has been adopted. Further, Decision Tree, Gaussian Naïve Bayes, K-Nearest neighbour, Logistic Regression, and Support Vector Machine are adopted as Machine Learning and Statistical Regression Techniques for training of software maintainability dataset. Results demonstrate that the proposed HPOSMP model gives better performance as compared to the base SMP models.

A multiple criteria decision analysis based approach to remove uncertainty in SMP models

Software has to be updated frequently to match the customer needs. If software maintainability is not given priority, it affects the software development life cycle and maintenance expenses, which deplete organizational assets. Before releasing software, maintainability must be estimated, as the impact of bugs and errors can affect the cost and reputation of the organization after deployment. Regardless of the programming paradigm, it’s important to assess software maintainability. Many software maintainability prediction models’ compatibilities with new programming paradigms are criticized because their limited applicability over heterogeneous datasets. Due this challenge small and medium-sized organizations may even skip the maintainability assessment, resulting in huge lose to such organizations. Motivated by this fact, we used Genetic Algorithm optimized Random Forest technique (GA) for software maintainability prediction models over heterogeneous datasets. To find optimal model for software maintainability prediction, the Technique for Order preference by Similarity to Ideal Solution (TOPSIS), a popular multiple-criteria decision-making model, is adopted. From the results, it is concluded that the GA is optimal for predicting maintainability of software developed in various paradigms.

Impact of Historical Software Metric Changes in Predicting Future Maintainability Trends in Open-Source Software Development

Software maintenance is one of the key stages in the software lifecycle and it includes a variety of activities that consume the significant portion of the costs of a software project. Previous research suggest that future software maintainability can be predicted, based on various source code aspects, but most of the research focuses on the prediction based on the present state of the code and ignores its history. While taking the history into account in software maintainability prediction seems intuitive, the research empirically testing this has not been done, and is the main goal of this paper. This paper empirically evaluates the contribution of historical measurements of the Chidamber & Kemerer (C&K) software metrics to software maintainability prediction models. The main contribution of the paper is the building of the prediction models with classification and regression trees and random forest learners in iterations by adding historical measurement data extracted from previous releases gradually. The maintainability prediction models were built based on software metric measurements obtained from real-world open-source software projects. The analysis of the results show that an additional amount of historical metric measurements contributes to the maintainability prediction. Additionally, the study evaluates the contribution of individual C&K software metrics on the performance of maintainability prediction models.

Using Ensembles for Class-Imbalance Problem to Predict Maintainability of Open Source Software

To facilitate software maintenance and save the maintenance cost, numerous machine learning (ML) techniques have been studied to predict the maintainability of software modules or classes. An abundant amount of effort has been put by the research community to develop software maintainability prediction (SMP) models by relating software metrics to the maintainability of modules or classes. When software classes demanding the high maintainability effort (HME) are less as compared to the low maintainability effort (LME) classes, the situation leads to imbalanced datasets for training the SMP models. The imbalanced class distribution in SMP datasets could be a dilemma for various ML techniques because, in the case of an imbalanced dataset, minority class instances are either misclassified by the ML techniques or get discarded as noise. The recent development in predictive modeling has ascertained that ensemble techniques can boost the performance of ML techniques by collating their predictions. Ensembles themselves do not solve the class-imbalance problem much. However, aggregation of ensemble techniques with the certain techniques to handle class-imbalance problem (e.g., data resampling) has led to several proposals in research. This paper evaluates the performance of ensembles for the class-imbalance in the domain of SMP. The ensembles for class-imbalance problem (ECIP) are the modification of ensembles which pre-process the imbalanced data using data resampling before the learning process. This study experimentally compares the performance of several ECIP using performance metrics Balance and g-Mean over eight Apache software datasets. The results of the study advocate that for imbalanced datasets, ECIP improves the performance of SMP models as compared to classic ensembles.

Comparative Analysis of Object-Oriented Software Maintainability Prediction Models

Abstract Software maintainability is one of the most important aspects when evaluating the quality of a software product. It is defined as the ease with which the existing software can be modified. In the literature, several researchers have proposed a large number of models to measure and predict maintainability throughout different phases of the Software Development Life Cycle. However, only a few attempts have been made for conducting a comparative study of the existent proposed prediction models. In this paper, we present a detailed classification and conduct a comparative analysis of Object-Oriented software maintainability prediction models. Furthermore, we considered the aforementioned proposed models from three perspectives, which are architecture, design and code levels. To the best of our knowledge, such an analysis that comprises the three levels has not been conducted in previous research. Moreover, this study hints at certain fundamental basics concerning the way of how measure the maintainability knowing that at each level the maintainability will be measured differently. In addition, we will focus on the strengths and weaknesses of these models. Consequently, the comparative study yields that several statistical and machine learning techniques have been employed for software maintainability prediction at code level during the last decade, and each technique possesses its specific characteristic to develop an accurate prediction model. At the design level, the majority of the prediction models measured maintainability according to the characteristics of the quality models. Whereas at the architectural level, the techniques adopted are still limited and only a few of studies have been conducted in this regard.

STATISTICAL COMPARISON OF MODELLING METHODS FOR SOFTWARE MAINTAINABILITY PREDICTION

The objective of this paper is statistical comparison of modelling methods for software maintainability prediction. The statistical comparison is performed by building software maintainability prediction models using 27 different regression and machine learning based algorithms. For this purpose, software metrics datasets of two different commercial object-oriented systems are used. These systems were developed using an object oriented programming language Ada. These systems are User Interface Management System (UIMS) and Quality Evaluation System (QUES). It is shown that different measures like MMRE, RMSE, Pred(0.25) and Pred(0.30) calculated on predicted values obtained from leave one out (LOO) cross validation produce very divergent results regarding accuracy of modelling methods. Therefore the 27 modelling methods are evaluated on the basis of statistical significance tests. The Friedman test is used to rank various modelling methods in terms of absolute residual error. Six out of the ten top ranked modelling methods are common to both UIMS and QUES. This indicates that modelling methods for software maintainability predicton are solid and scalable. After obtaining ranks, pair wise Wilcoxon Signed rank test is performed. Wilcoxon Sign rank test indicates that the top ranking method in UIMS data set is significantly superior to only four other modelling methods whereas the top tanking method in QUES data set is significantly superior to 11 other modelling methods. The performance of instance based learning algorithms — IBk and Kstar is comparable to modelling methods used in earlier studies.

Predicting object-oriented software maintainability using multivariate adaptive regression splines

Accurate software metrics-based maintainability prediction can not only enable developers to better identify the determinants of software quality and thus help them improve design or coding, it can also provide managers with useful information to help them plan the use of valuable resources. In this paper, we employ a novel exploratory modeling technique, multiple adaptive regression splines (MARS), to build software maintainability prediction models using the metric data collected from two different object-oriented systems. The prediction accuracy of the MARS models are evaluated and compared using multivariate linear regression models, artificial neural network models, regression tree models, and support vector models. The results suggest that for one system MARS can predict maintainability more accurately than the other four typical modeling techniques, and that for the other system MARS is as accurate as the best modeling technique.