Software Maintainability Prediction Research Articles

Software maintainability prediction using an enhanced random forest algorithm

Now-a-days, software maintainability has become one of the significant quality-assessing attributes for any software system. Current study presents an enhanced-RFA (Random Forest Algorithm) approach for Software Maintainability Prediction. The proposed approach combines Random Forest (RF) algorithm with three prevalent feature selection techniques namely Chi-Squared, RF and Linear Correlation Filter along with a re-sampling technique intended to improve the prediction accuracy of basic RF algorithm. Enhanced- RFA is applied on two commercially available datasets, QUES and UIMS and performance is evaluated on the basis of R2. Results indicate that the proposed approach performs significantly better than RFA for both the datasets with an improvement in R2 values equal to 69.50%, 65.57% & 69.40% for QUES and 31.90%, 44.94% & 51.81% for UIMS using chisquared, RF and linear correlation filter techniques respectively.

Assessing Cross-Project Technique for Software Maintainability Prediction

Software Maintainability refers to the ease with which software maintenance activities like correction of faults, deletion of obsolete code, addition of new code etc. can be carried out to adapt to the modified environment. Predicting maintainability in early stages of development helps in reducing the cost of maintenance and ensures optimum utilization of resources. Sometimes, it becomes difficult to train prediction models using historical data of the same dataset for which the model is being developed because of the unavailability of sufficient amount of training data, in turn making a way for Cross-Project technique for Software Maintainability Prediction (CPSMP). In order to evaluate the proposed CPSMP technique, QUES dataset is used as training set and UIMS dataset is used as test set in this study with 19 different regression modelling methods. Performance of CPSMP model is evaluated using Root Mean Square Error (RMSE) as an accuracy measure. Results show that cross-project technique can successfully be applied for maintainability prediction. The average RMSE value calculated for all the modelling methods is found to be 82.310 without CPSMP whereas an average RMSE value of 71.532 is obtained with CPSMP resulting in an overall improvement in prediction performance by 13.09%. Also, 84.21% of the total techniques used in this study performed better with CPSMP.

A systematic literature review of machine learning techniques for software maintainability prediction

Abstract Context Software maintainability is one of the fundamental quality attributes of software engineering. The accurate prediction of software maintainability is a significant challenge for the effective management of the software maintenance process. Objective The major aim of this paper is to present a systematic review of studies related to the prediction of maintainability of object-oriented software systems using machine learning techniques. This review identifies and investigates a number of research questions to comprehensively summarize, analyse and discuss various viewpoints concerning software maintainability measurements, metrics, datasets, evaluation measures, individual models and ensemble models. Method The review uses the standard systematic literature review method applied to the most common computer science digital database libraries from January 1991 to July 2018. Results We survey 56 relevant studies in 35 journals and 21 conference proceedings. The results indicate that there is relatively little activity in the area of software maintainability prediction compared with other software quality attributes. CHANGE maintenance effort and the maintainability index were the most commonly used software measurements (dependent variables) employed in the selected primary studies, and most made use of class-level product metrics as the independent variables. Several private datasets were used in the selected studies, and there is a growing demand to publish datasets publicly. Most studies focused on regression problems and performed k-fold cross-validation. Individual prediction models were employed in the majority of studies, while ensemble models relatively rarely. Conclusion Based on the findings obtained in this systematic literature review, ensemble models demonstrated increased accuracy prediction over individual models, and have been shown to be useful models in predicting software maintainability. However, their application is relatively rare and there is a need to apply these, and other models to an extensive variety of datasets with the aim of improving the accuracy and consistency of results.

Fuzzy Network Based Framework for Software Maintainability Prediction

Software metrics based maintainability prediction is leading to development of new sophisticated techniques to construct prediction models. This paper proposes a new software maintainability prediction framework, which bases on Fuzzy Network, a novel exploratory modeling technique. The proposed framework utilizes both the metric data collected from software system and the subjective appraisals from experts. An application example of the framework is shown. In comparison to the Standard Fuzzy System based models, Fuzzy Network based models improves the transparency more than 71.3% and the accuracy more than 11.0%. It is confirmed that Fuzzy Network based framework is more appropriate for constructing SMP model.

Optimized neuro-PSO-based software maintainability prediction using relief features selection method

<p>The recent development in software engineering reveals the importance of software maintenance during the time of software development that is becoming more important in software development environment and software metrics, which are very essential for measuring the maintainability of software, software complexity, estimating size, quality and project efforts. There are various approaches through which one can estimate the software cost and predict on various kinds of deliverable items. This paper aims at developing an optimized Neuro-PSO-based software maintainability prediction model by applying the dimensionality reduction using relief feature selection method for identifying the optimal feature subsets in order to increase the accuracy and reduce the time complexity of the prediction model. The simulation result proves the performance of the proposed model which will be more beneficial for the software developers in predicting the maintenance of the software in advance.</p>

Estimation of maintainability parameters for object-oriented software using hybrid neural network and class level metrics

The various software metrics proposed in the literature can be used to evaluate the quality of software systems written in object-oriented manner. These metrics are broadly categorized into two subcategories i.e., system level software metrics and class level software metrics. In this work, ten different types of class level metrics are considered as an input to develop one model for predicting software maintainability of object-oriented software system. These models are developed using three types of neural networks, i.e., artificial neural network, radial basis function network, and functional link artificial neural network. In this study, a hybrid algorithm based on genetic algorithm (GA) with gradient descent algorithm has been proposed to find optimal weights of these neural networks. Since accuracy of the prediction model is highly dependent on the class level metrics, they are considered as input of the models. So, five different feature selection techniques are used in this study to identify the best set of features with an objective to improve the accuracy of software maintainability prediction model. The effectiveness of these models are evaluated using four evaluation metrics, i.e., MAE, MMRE, RMSE, and SEM. In this work, parallel computing concept has been also considered with an objective to reduce the model training time. The results show that the model developed using the proposed hybrid algorithm based on GA with gradient descent algorithm give better results as compared to the work presented by other authors in literature. The results also show that feature selection techniques obtain better results for predicting maintainability as compared to all metrics. The experimental results show that parallel computing is beneficial in reducing the model training time.

Deep Learning Approach for Software Maintainability Metrics Prediction

Software maintainability predicts changes or failures that may occur in software after it has been deployed. Since it deals with the degree to which an application may be understood, repaired, or enhanced, it also takes into account the overall cost of the project. In the past, several measures have been taken into account for predicting metrics that influence software maintainability. However, deep learning is yet to be explored for the same. In this paper, we perform deep learning for software maintainability metrics' prediction on a large number of datasets. Unlike the previous research works, we have relied on large datasets from 299 software and subsequently applied various metrics and functions to the same; 29 object-oriented metrics have been considered along with their impact on software maintainability of open source software. Several metrics have been analyzed and descriptive statistics of these metrics have been pointed out. The proposed long short term memory has been evaluated using measures, such as mean absolute error, root mean square error and accuracy. Five machine learning algorithms, namely, ridge regression with variable selection, decision tree, quantile regression forest, support vector machine, and principal component analysis have been applied to the original datasets, as well as, to the refined datasets. It was found that this paper provides results in the form of metrics that may be used in the prediction of software maintenance and the proposed deep learning model outperforms all of the other methods that were considered. Furthermore, the results of experiment affirm the efficiency of the proposed deep learning model for software maintainability prediction.

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.

Performance of Maintainability Index prediction models: a feature selection based study

Numerous design metrics have been studied in the literature to assess software performance and future maintainability issues. Maintainability Index (MI) is an indicator of software maintenance efforts, but that can be computed only after the complete code is developed. MI is complex to calculate and hence has been estimated by several researchers using design metrics. Various prediction models are built using different sets of design metrics as independent variables. Further, researchers have used different prediction models for predicting software maintainability. To identify most prominent design metrics and the prediction models that show consistent performance irrespective of the choice of design metrics, the present empirical study is performed on datasets of 26 open source Java projects with more than one thousand and six hundred class hierarchies. Seven feature selection methods are applied to find out the most prominent metrics that are selected by majority of methods. Then four most frequently used prediction models namely, multi linear regression, multilayer perceptron, support vector regression and M5P regression tree are analyzed for their performance with respect to initial selection of design metrics. The study concludes that MIF, MaxDIT, LCC, TCC are the most significant predictors for Maintainability Index of class hierarchies. Among the all the prediction models, support vector regression model exhibits the best performance both with respect to the choice of metrics and irrespective of initial metric selection.

Software maintainability prediction using hybrid neural network and fuzzy logic approach with parallel computing concept

In present day scenario, majority of software companies use object-oriented concept to develop software systems as it enables effective design, development, testing and maintenance, in addition to the optimal characterization of the software system. With the increase in number of these software systems, their effective maintenance aspect becomes very important day by day. In this study, Neuro-Fuzzy approach: hybrid neural network and fuzzy logic approach has been considered to develop a maintainability model using ten different object-oriented static source code metrics as input. This method is applied on maintainability data of two commercial software products such as UIMS and QUES. Rough set analysis (RSA) and principal component analysis (PCA) are used to select suitable set of metrics from the ten metrics employed to improve performance of maintainability prediction model. From experimental results, it is observed that Neuro-Fuzzy model can effectively predict the maintainability of object-oriented software systems. After implementing parallel computing concept, it is observed that the training time gets reduced to a significant amount when the number of computing nodes were increased. Further it is observed that selected subset of metrics using feature selection techniques i.e., PCA, and RSA was able to predict maintainability with higher accuracy.

Three empirical studies on predicting software maintainability using ensemble methods

More accurate prediction of software maintenance effort contributes to better management and control of software maintenance. Several research studies have recently investigated the use of computational intelligence models for software maintainability prediction. The performance of these models, however, may vary from dataset to dataset. Consequently, ensemble methods have become increasingly popular as they take advantage of the capabilities of their constituent computational intelligence models toward a dataset to come up with more accurate or at least competitive prediction accuracy compared to individual models. This paper investigates and empirically evaluates different homogenous and heterogeneous ensemble methods in predicting software maintenance effort and change proneness. Three major empirical studies were designed and conducted taken into consideration different design such as the types of the investigated ensembles methods, types of prediction problems, used datasets, and other experimental setup. Overall empirical evidence obtained from the three studies confirms that some ensemble methods provide more accurate or at least competitive prediction accuracy compared to individual models across datasets, and thus they are more reliable.

SMPLearner: learning to predict software maintainability

Accurate and practical software maintainability prediction enables organizations to effectively manage their maintenance resources and guide maintenance-related decision making. This paper presents SMPLearner, an automated learning-based approach to train maintainability predictors by harvesting the actual average maintenance effort computed from the code change history as well as employing a much richer set of 44 four-level hierarchical code metrics collected by static code analysis tools. We systematically evaluated SMPLearner on 150 observations partitioned from releases of eight large scale open source software systems. Our evaluation showed that SMPLearner not only outperformed the traditional 4-metric MI model but also the recent learning-based maintainability predictors constructed based on single Class-level metrics, demonstrating that single Class-level metrics were not sufficient for maintainability prediction.

Fuzzy Based Approach for Predicting Software Maintainability

Software maintenance is a process of modifying existing operational software while leaving its primary functions intact. Software maintenance encompasses a broad range of activates, like error correction, enhancement of capabilities, deletion of obsolete capabilities and optimization.sofwtare maintainability assessment is major issue these days. producing software that is easy to maintain may save large costs in industries. the maintenance of existing software can account for 70% of the total efforts putin application development[Pres05].the value of software can be enhanced by meeting additional requirements, making it easier to use, improving efficiency and employing newer technologies. this paper discusses various issues and challenges, related with the maintainability assessment of software systems. The present work proposes a fuzzy logic based approach for quantification of maintainability of software system based on combined effect of four major aspects of software. i.e average number of live variables ,average life span of variables, average cyclomatic complexity and the comment ratio. Classroom projects are considered to estimate and validate the proposed maintainability model. Keyword: Cylomatic complexity, comment ratio, maintainability, triangular fuzzy number 1. METRICS FOR MAINTAINABILITY ASSESSMENT Researcher have tried to quantify maintainability in different types of measures[12,13,14].here we consider four major aspects of software to assess maintainability,i.e average number of live variables, average life span of variables, average cyclomatic complexity and comment ratio. a) AVERAGE NUMBER OF LIVE VARIABLES(ALV): a variable is live at a particular statement only if a certain number of statements reference it before or after that statement. the average number of live variable is the sum of the count of live variables divided by the count of executable statements. the higher, the average number of live variables, the more difficult it would be to develop and to maintain the software. b) THE AVERAGE LIFE SPAN OF VARIABLES(ALS): The life span of a variable is defined as the number of statements between two successive references of the same variable. The average life span of variable is the Ratio of the sum of life spans to number of variables. C) COMMENT RATIO(CR): The comment ratio is defined as CR=(S+C)/C------------------(1) Where S denotes total line of code C represents total number of comment lines. International Journal of Engineering Trends and Technology (IJETT) – Volume 10 Number 1 Apr 2014 ISSN: 2231-5381 http://www.ijettjournal.org Page 9 A lower comment ratio means better readability and the better maintainability. D) AVERAGE CYCLOMATIC COMPLEXITY (ACC): The average cyclomatic complexity is defined as the average cyclomatic complexity of all modules, where cyclomatic complexity is defined as V=e-n+2p----------------------(2) Where e is number of edges in a program flow graph,n is the number of nodes, and p is the number of connected Components. Maintainability of software declines with increasing average cyclomatic complexity. 2. FUZZY SET THEORY FOR MAINTAINABILITY: Recently many researchers have proposed some integrated models for maintainability measurement, which leave significant room for further improvements. here we proposed a methodology to improve the maintainability metrics system based on the fuzzy set theory.ALV,ALS,ACC and CR are classified in four levels of complexity i.e low, medium high and very high and corresponding weights are then assigned for each. The complexity levels and their corresponding weights for ALV,ALS ACC and CRR are described in table 1 and 2 TABLE1: COMPLEXITY AND WEIGHTS FOR ALV ALV Complexity weight 0-2 Low W1 2-5 Medium W2 5-8 High W3 8 or more Very High W4 TABLE2: COMPLEXITY AND WEIGHT FOR ALS ALS complexity Weight 0-20 Low W1 20-150 Medium W2 150-400 High W3 400 or more Very high W4 TABLE4: COMPLEXITY AND WEIGHT FOR ACC ACC complexity Weight 0-5 Low W1 5-13 Medium W2 13-18 High W3 18 or more Very high W4 TABLE5: COMPLEXITY AND WEIGHT FOR CR ALS complexity Weight 0-4 Low W1 4-8 Medium W2 8-12 High W3 12 or more Very high W4 The set w1,w2,w3 and w4 may be different for program in different languages . 3. FUZZIFICATION: The complexity attributes low, medium and high for the four metrics ALV,ALS,ACC,and CR are taken as triangular fuzzy number(TFN).membership functions are evaluated using complexity and coefficient matrices. fuzzy pictorial representation of TFNs for ALV,ALS ACC and CR are shown in Figure: Fig1.Fuzzy Pictorial Representation of ALV International Journal of Engineering Trends and Technology (IJETT) – Volume 10 Number 1 Apr 2014 ISSN: 2231-5381 http://www.ijettjournal.org Page 10 Fig2.Fuzzy Pictorial Representation of ALS Fig3.Fuzzy Pictorial Representation of ACC Fig4.Fuzzy Pictorial Representation of CR 4. Defuzzification: Defuzzification process is applied to evaluate crisp value of maintainability factors for ALV,ALS,ACC and CRR.Ralv Racc and Rcr are used to represent corresponding factors of maintainability for ALV,ALS,ACC and CR repectively.Defuzzification rules are defined in the following equations.

Application of Group Method of Data Handling model for software maintainability prediction using object oriented systems

Object-oriented methodology has emerged as most prominent in software industry for application development. Maintenance phase begins once the product is delivered and by software maintainability we mean the ease with which existing software could be modified during maintenance phase. We can improve and control software maintainability if we can predict it in the early phases of software life cycle using design metrics. Predicting the maintainability of any software has become critical with the increasing importance of software maintenance. Many authors have practiced and proved theoretical validation followed by empirical evaluation using statistical and experimental techniques for evaluating the relevance of any given metrics suite using many models. In this paper, we have presented an empirical study to evaluate the effectiveness of novel technique called Group Method of Data Handling (GMDH) for the prediction of maintainability over other models. Although many metrics have been proposed in the literature, software design metrics suite proposed by Chidamber et al. and revised by Li et al. have been selected for this study. Two web-based customized softwares developed using C# Language have been used for empirical study. Source code of old and new versions for both applications were collected and analysed against modifications made in every class. The changes were counted in terms of number of lines added, deleted or modified in the classes belonging to new version with respect to the classes of old version. Finally values of metrics were combined with “change” in order to generate data points. Hence, in this study an attempt has been made to evaluate and examine the effectiveness of prediction models for the purpose of software maintainability using real life web based projects. Three models using Feed Forward 3-Layer Back Propagation Network (FF3LBPN), General Regression Neural Network (GRNN) and GMDH are developed and performance of GMDH is compared against two others i.e. FF3LBPN and GRNN. With the aid of this empirical analysis, we can safely suggest that software professionals can use OO metric suite to predict the maintainability of software using GMDH technique with least error and best precision in an object oriented paradigm.

Machine learning approaches for predicting software maintainability: a fuzzy‐based transparent model

Software quality is one of the most important factors for assessing the global competitive position of any software company. Thus, the quantification of the quality parameters and integrating them into the quality models is very essential.Many attempts have been made to precisely quantify the software quality parameters using various models such as Boehm's Model, McCall's Model and ISO/IEC 9126 Quality Model. A major challenge, although, is that effective quality models should consider two types of knowledge: imprecise linguistic knowledge from the experts and precise numerical knowledge from historical data.Incorporating the experts’ knowledge poses a constraint on the quality model; the model has to be transparent.In this study, the authorspropose a process for developing fuzzy logic-based transparent quality prediction models.They applied the process to a case study where Mamdani fuzzy inference engine is used to predict software maintainability.Theycompared the Mamdani-based model with other machine learning approaches.The resultsshow that the Mamdani-based model is superior to all.

Machine Learning Algorithm Evaluate the maintainability

Maintainability is an important quality attribute and a difficult concept as it involves a number of measurements. Quality estimation means estimating maintainability of software. Maintainability is a set of attribute that bear on the effort needed to make specified modification. The main goal of this paper is to propose use of few machine learning algorithms with an objective to predict software maintainability and evaluate them. The propose models are Gaussian process regression networks (GPRN), probably approximately correct learning (PAC), Genetic algorithm (GA). This paper predicts the maintenance effort. The QUES (Quality evaluation system) dataset are used in this study. The QUES datasets contains 71 classes. To measure the maintainability, number of “CHANGE” is observed over a period of few years. We can define CHANGE as the number of lines of code which were added, deleted or modified during few year maintenance periods. After this study these machine learning algorithm was compared with few models such as GRNN (General regression neural network) model, RT (Regression tree), MARS (Multiple adaptive regression
 splines), SVM (Support vector machine), MLR (Multiple linear regression) models. Based on experiments, it was
 found that GPRN can be predicting the maintainability more accurately and precisely than prevailing models. We also include object oriented software metric to measure the software maintainability. The use of machine learning algorithms to establish the relationship between metrics and maintainability would be much better approach as these are based on quantity as well as quality.

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.

Support Vector approach by using Radial Kernel Function for Prediction of Software Maintenance Effort on the basis of Multivariate Approach

The connection among object oriented metrics and software maintenance effort is complex and non-linear. Therefore, there is wide research area in development and application of sophisticated techniques which can be used to construct models for predicting software maintenance effort. The aim of this paper is to evaluate Support Vector Machine for Regression in software maintainability prediction using object-oriented metrics to construct models for prediction of Software Maintenance Effort. Support Vector Machine has already proved its importance in Banking Sector and in other areas also. I am using SVR with Radial kernel function. It is observed that Support Vector Machine can be used for constructing accurate models for prediction of software maintenance effort which gives most accurate models for prediction. We are using maintenance effort data of two commercial software products QUES (Quality Evaluation System) and UIMS (User Interface System) Data is used in this study. The dependent variable in our study is maintenance effort. The independent variables are eight Object Oriented metrics. I will verify the dataset by Multivariate performance basis.

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.

Construction and testing of polynomials predicting software maintainability

Assessing software maintainability is complicated by the many aspects of software and software systems that affect maintenance activity. In an earlier study of 35 published works on software maintainability, we coalesced software maintainability definitions into a hierarchical structure containing 92 known maintainability attributes (Oman et al., 1991). The tree structure is refined through successive subtrees until a leaf node, representing an identified maintainability attribute, is defined. Our hierarchy serves as a taxonomic definition for software maintainability that is compatible with the 35 published works on which it is based. In a subsequent study, we showed how these software maintainability attributes can be measured via a battery of metrics and discussed methods to create a single value (hybrid metric) that represents the maintainability of the software system (Oman and Hagemeister, 1992). We were charged with identifying the smallest set of metrics useful in predicting the maintainability of a suite of software modules.’ In this article, we show how the set of applicable measures can be reduced to the minimum number of metrics useful in predicting maintainability. The maintainability assessment polynomials presented here are accurate models of the test data on which they were constructed. We do not claim that these are the only models for predicting maintainability, nor do we claim they are the best overall, but they are automated working models that can be used to quickly and easily predict software maintainability. The models presented here have been used in conjunction with two prototype software tools developed at the University of Idaho: Hewlett-Packard Software Maintainability Estimation Spreadsheet