Heterogeneous Software Defect Prediction using Generative Models

Heterogeneous Software Defect Prediction using Generative Models

It is widely known that one of the most helpful phases in SDLC is Software Defect Prediction (SDP). SDP can tell which modules have the highest chances to have bugs and companies can then devote suitable time in analysing those modules. SDP can be extremely helpful but it can be very difficult to pr...

Full description

Saved in:
Bibliographic Details
Published in:2022 IEEE 11th International Conference on Communication Systems and Network Technologies (CSNT) pp. 367 - 372
Main Authors: Jain, Bhavye, Patidar, Sanjay, Sudershan, Divya
Format: Conference Proceeding
Language:English
Published: IEEE 23-04-2022
Subjects:
Autoencoder
Communication systems
Companies
Computer bugs
Conferences
Cross Project
Datasets
Defect Prediction
Generative adversarial networks
Predictive models
Static VAr compensators
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is widely known that one of the most helpful phases in SDLC is Software Defect Prediction (SDP). SDP can tell which modules have the highest chances to have bugs and companies can then devote suitable time in analysing those modules. SDP can be extremely helpful but it can be very difficult to predict buggy modules. Recently many studies have shown how in cases where enough data is not present, defect data from different projects can be used; this is called cross-project defect prediction (CPDP). However most of these studies require Homogeneous data for prediction i.e. our source dataset and target dataset must have the same parameters. Many times, especially for newer startups, previous data is not present for defect prediction, in such a case we want to be able to predict defects with data from other projects. We have trained and compared the performance of three models namely a GAN, Variational Autoencoder, and Adversarial Autoencoder for quickly transferring lessons learned about defect prediction among different datasets.
DOI:10.1109/CSNT54456.2022.9787607