Open Access

ARTICLE

Modeling Metaheuristic Optimization with Deep Learning Software Bug Prediction Model

by M. Sangeetha^1,*, S. Malathi²

1 Department of Computer Science and Engineering, Sathyabama Institute of Science and Technology, Chennai, 600 0119, India
2 Department of Computer Science and Engineering, Panimalar Engineering College, Chennai, 600 123, India

* Corresponding Author: M. Sangeetha. Email:

Intelligent Automation & Soft Computing 2022, 34(3), 1587-1601. https://doi.org/10.32604/iasc.2022.025192

Received 15 November 2021; Accepted 09 February 2022; Issue published 25 May 2022

Abstract

Software testing is an effective means of verifying software stability and trustworthiness. It is essential in the software development process and needs a huge quantity of resources such as labor, money, and time. Automated software testing can be used to save manual work, shorten testing times, and improve testing performance. Recently, Software Bug Prediction (SBP) models have been developed to improve the software quality assurance (SQA) process through the prediction of bug parts. Advanced deep learning (DL) models can be used to classify faults in software parts. Because hyperparameters have a significant impact on the performance of any DL model, a proper hyperparameter optimization approach utilizing metaheuristic methods is required. This paper provides a unique Metaheuristic Optimization with Deep Learning based SBP (MODL-SBP) methodology to ensure software dependability and trustworthiness. The suggested technique entails creating a hybrid Convolution Neural Network (CNN) bi-directional long short-term memory (BiLSTM) to forecast software problems. Furthermore, the Chaotic Quantum Grasshopper Optimization Algorithm (CQGOA) is used for hyperparameter optimization of the CNN-BiLSTM models, which enhances predictive accuracy. To demonstrate the superior performance of the MODL-SBP technique, a wide range of simulations are performed on benchmark datasets, with the results highlighting the superior performance of the proposed model over other recent techniques.

---

Keywords

---