A Review of Cross-Project and Just-in-Time Software Defect Prediction with Lessons from Real-World Failures

Software defects that escape into production can ground airline fleets, halt trading desks and disrupt essential services within minutes. This review connects the research on software defect prediction with the real cost of such failures. It first examines a set of widely reported incidents in aviation, airlines, finance and banking, drawn from official investigation reports, company filings and news coverage, and summarises their causes and documented losses. It then reviews two lines of defect prediction research that matter most in industrial settings. Cross-project defect prediction (CPDP) addresses the common situation in which a new or small project has too little history to train its own model, and must borrow knowledge from other projects through instance selection, feature transformation, transfer learning and deep domain adaptation. Just-in-time defect prediction (JIT-DP) scores each code change at the moment it is committed, so that risky changes can be reviewed before they reach production; deep learning models for JIT-DP learn directly from commit messages and code diffs. The review compares methods, datasets and evaluation practices in both lines, and then asks a practical question: for each incident, at which stage could a prediction model plausibly have raised a warning, and where were the causes outside its reach? It closes with a research agenda for deployment-aware defect prediction that combines cross-project knowledge, change-level risk scores and release-time safeguards.

Authors

Publication Details

Journal
International Journal of Innovative Science and Research Technology (IJISRT)
Published
2026-11-21
DOI
https://doi.org/10.38124/ijisrt/26nov573
Primary Topic
Software Engineering Research
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Review of Cross-Project and Just-in-Time Software Defect Prediction with Lessons from Real-World Failures

Radha Lakshmi Rajagopalan, Radha V.
International Journal of Innovative Science and Research Technology (IJISRT)
Software Engineering Research
article

A Review of Cross-Project and Just-in-Time Software Defect Prediction with Lessons from Real-World Failures

Radha Lakshmi Rajagopalan, Radha V.
article en

Abstract

Software defects that escape into production can ground airline fleets, halt trading desks and disrupt essential services within minutes. This review connects the research on software defect prediction with the real cost of such failures. It first examines a set of widely reported incidents in aviation, airlines, finance and banking, drawn from official investigation reports, company filings and news coverage, and summarises their causes and documented losses. It then reviews two lines of defect prediction research that matter most in industrial settings. Cross-project defect prediction (CPDP) addresses the common situation in which a new or small project has too little history to train its own model, and must borrow knowledge from other projects through instance selection, feature transformation, transfer learning and deep domain adaptation. Just-in-time defect prediction (JIT-DP) scores each code change at the moment it is committed, so that risky changes can be reviewed before they reach production; deep learning models for JIT-DP learn directly from commit messages and code diffs. The review compares methods, datasets and evaluation practices in both lines, and then asks a practical question: for each incident, at which stage could a prediction model plausibly have raised a warning, and where were the causes outside its reach? It closes with a research agenda for deployment-aware defect prediction that combines cross-project knowledge, change-level risk scores and release-time safeguards.

International Journal of Innovative Science and Research Technology (IJISRT)Vol. 11(11)
Software Engineering Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
A Review of Cross-Project and Just-in-Time Software Defect Prediction with Lessons from Real-World Failures — Radha Lakshmi Rajagopalan, Radha V. · International Journal of Innovative Science and Research Technology (IJISRT) (2026) | TGRS Research Map | TGRS