An Improved and Optimized Random Forest Based Approach to Predict the Software Faults
Open Access
- 9 May 2024
- journal article
- research article
- Published by Springer Nature in SN Computer Science
- Vol. 5 (5) , 1-18
Abstract
Effective software fault prediction is crucial for minimizing errors during software development and preventing subsequent failures. This research introduces an enhanced Random Forest-based approach for predicting software faults, specifically focusing on the NASA JM1 dataset. The dataset comprises 21 software metrics indicating the presence or absence of faults in a module, and it is utilized to evaluate the proposed approach. The study delves into the intricacies of the NASA dataset, detailing the cleaning process and addressing class imbalance through Synthetic Minority Over-sampling Technique (SMOTE). The core of our approach involves the implementation and fine-tuning of the Random Forest classifier, with a specific focus on optimizing hyperparameters to enhance predictive accuracy. In comparative evaluations with standard machine learning models, our proposed approach demonstrated superior performance, achieving an accuracy of 82.96% and an F1 score of 89.53%. Notably, we emphasize the significance of software defects and their potential to cause failures and crashes during software development, leading to substantial organizational losses. The paper provides a comprehensive examination of different aspects of the machine learning model, offering detailed insights, examples, and illustrative figures to enhance the understanding of our proposed approach.Keywords
Funding Information
- Manipal Academy of Higher Education, Manipal
This publication has 18 references indexed in Scilit:
- A Classification Model for Software Bug Prediction Based on Ensemble Deep Learning Approach Boosted with SMOTE Technique2021
- Hyper-Parameter Optimization of Classifiers, Using an Artificial Immune Network and Its Application to Software Bug PredictionIEEE Access, 2020
- Fine-grained just-in-time defect predictionJournal of Systems and Software, 2019
- Hyperparameters and tuning strategies for random forestWIREs Data Mining and Knowledge Discovery, 2019
- Path testing based reliability analysis framework of component based software systemMeasurement, 2018
- Software bug prediction using weighted majority voting techniquesAlexandria Engineering Journal, 2018
- Evaluating the Effectiveness of Conventional Machine Learning Techniques for Defect Prediction: A Comparative Study2018
- Evaluating Complexity, Code Churn, and Developer Activity Metrics as Indicators of Software VulnerabilitiesIEEE Transactions on Software Engineering, 2010
- A critique of software defect prediction modelsIEEE Transactions on Software Engineering, 1999
- Empirically guided software development using metric-based classification treesIEEE Software, 1990