A three-stage transfer learning framework for multi-source cross-project software defect prediction
Bibliographic record
Abstract
Transfer learning techniques have been proved to be effective in the field of Cross-project defect prediction (CPDP). However, some questions still remain. First, the conditional distribution difference between source and target projects has not been considered. Second, facing multiple source projects, most studies only rarely consider the issues of source selection and multi-source data utilization; instead, they use all available projects and merge multi-source data together to obtain one final dataset. To address these issues, in this paper, we propose a three-stage weighting framework for multi-source transfer learning (3SW-MSTL) in CPDP. In stage 1, a source selection strategy is needed to select a suitable number of source projects from all available projects. In stage 2, a transfer technique is applied to minimize marginal differences. In stage 3, a multi-source data utilization scheme that uses conditional distribution information is needed to help guide researchers in the use of multi-source transferred data. First, we have designed five source selection strategies and four multi-source utilization schemes and chosen the best one to be used in stage 1 and 3 in 3SW-MSTL by comparing their influences on prediction performance. Second, to validate the performance of 3SW-MSTL, we compared it with four multi-source and six single-source CPDP methods, a baseline within-project defect prediction (WPDP) method, and two unsupervised methods on the data from 30 widely used open-source projects. Through experiments, bellwether and weighted vote are separately chosen as a source selection strategy and a multi-source utilization scheme used in 3SW-MSTL. And, our results indicate that 3SW-MSTL outperforms four multi-source, six single-source CPDP methods and two unsupervised methods. And, 3SW-MSTL is comparable to the WPDP method. The proposed 3SW-MSTL model is more effective for considering the two issues mentioned before.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.008 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".