Browsing by Author "Kumar, Sandeep"
Now showing 1 - 3 of 3
- Results Per Page
- Sort Options
Article Citation Count: 41Empirical analysis of change metrics for software fault prediction(Pergamon-elsevier Science Ltd, 2018) Mıshra, Alok; Kumar, Sandeep; Kumar, Kuldeep; Mishra, Alok; Catal, Cagatay; Software EngineeringA quality assurance activity, known as software fault prediction, can reduce development costs arid improve software quality. The objective of this study is to investigate change metrics in conjunction with code metrics to improve the performance of fault prediction models. Experimental studies are performed on different versions of Eclipse projects and change metrics are extracted from the GIT repositories. In addition to the existing change metrics, several new change metrics are defined and collected from the Eclipse project repository. Machine learning algorithms are applied in conjunction with the change and source code metrics to build fault prediction models. The classification model with new change metrics performs better than the models using existing change metrics. In this work, the experimental results demonstrate that change metrics have a positive impact on the performance of fault prediction models, and high-performance models can be built with several change metrics. (C) 2018 Elsevier Ltd. All rights reserved.Article Citation Count: 3A Paired Learner-Based Approach for Concept Drift Detection and Adaptation in Software Defect Prediction(Mdpi, 2021) Mıshra, Alok; Kumar, Sandeep; Mishra, Alok; Software EngineeringThe early and accurate prediction of defects helps in testing software and therefore leads to an overall higher-quality product. Due to drift in software defect data, prediction model performances may degrade over time. Very few earlier works have investigated the significance of concept drift (CD) in software-defect prediction (SDP). Their results have shown that CD is present in software defect data and tha it has a significant impact on the performance of defect prediction. Motivated from this observation, this paper presents a paired learner-based drift detection and adaptation approach in SDP that dynamically adapts the varying concepts by updating one of the learners in pair. For a given defect dataset, a subset of data modules is analyzed at a time by both learners based on their learning experience from the past. A difference in accuracies of the two is used to detect drift in the data. We perform an evaluation of the presented study using defect datasets collected from the SEACraft and PROMISE data repositories. The experimentation results show that the presented approach successfully detects the concept drift points and performs better compared to existing methods, as is evident from the comparative analysis performed using various performance parameters such as number of drift points, ROC-AUC score, accuracy, and statistical analysis using Wilcoxon signed rank test.Article Citation Count: 7Predicting reliability of software in industrial systems using a Petri net based approach: A case study on a safety system used in nuclear power plant(Elsevier, 2022) Mıshra, Alok; Sumit; Kumar, Sandeep; Singh, Lalit Kumar; Mishra, Alok; Software EngineeringContext: Software reliability prediction in the early stages of development can be propitious in many ways. The combinatorial models used to predict reliability using architectures such as fault trees, binary decision diagrams, etc. have limitations in modeling complex system behavior. On the other hand, state-based models such as Markov chains suffer from the state-space explosion problem, and they need transition probability among different system states to measure reliability. These probabilities are usually assumed or are obtained from the operational profile for which the system should be used in the field. Objective: The objective of this paper is to present a method for predicting the reliability of software in industrial systems using a generalized stochastic Petri nets based approach. The key idea is to violate the assumption of state transition probabilities in the Markov chain. The state transition probabilities are calculated using Petri net transitions' throughput by performing stationary analysis under the consideration to identify and handle dead markings in the Petri net. Method: Initially, a generalized stochastic Petri net of the system under consideration is generated from the standard system's specification. Thereafter, dead markings are identified in the Petri net which are further removed to perform steady-state analysis. At last, a Markov model is generated based on the reachability graph of the Petri net, which is further used to predict the system reliability. Results: The presented method has been applied to a safety-critical system, Shut Down System-1, of a nuclear power plant, which is operational in the Canada Deuterium Uranium reactor. The predicted reliability of the system using this method is 99.99966% which has been validated using the specified system requirements. To further validate and generalize the results, sensitivity analysis is performed by varying different system parameters. Conclusions: The method discussed in this paper presents a step of performing structural analysis on the Petri net of the system under consideration to identify and handle dead markings on the Petri net. It further handles the issue of assuming transition probabilities among the system states by calculating them using Petri net transitions' throughput.