Pith. sign in

REVIEW 1 major objections 49 references

LinkRank: A Learning-to-Rank Framework for One-to-Many Issue-Commit Traceability

T0 review · 1 major / 0 minor · reviewed 2026-07-30 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read One-to-many issue–commit traceability works better as issue-level ranking and iterative set selection than as independent pairwise classification.

desk verdict Solid MSR methods paper: set-recovery framing plus a real multi-commit dataset beat pairwise/LLM baselines under their protocol, but RDS pools are GT-time-anchored so the gap size needs a blind-pool check. read the letter →

arxiv 2607.23610 v1 pith:QT5VEFLB submitted 2026-07-26 cs.SE

Abhishek Kumar , Tuhin Mondal , Partha Pratim Das , Partha Pratim Chakrabarti This is my paper · ORCID
classification cs.SE
keywords issue–committraceabilitylinkrecoverylearningtoranksoftwaremaintenanceminingrepositoriesone-to-manylinkingiterativeselection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Software projects often fix a single issue with several commits, yet most automated link-recovery tools still treat each issue–commit pair in isolation and assume one commit per issue. This paper argues that the right task is set recovery: for each issue, rank a pool of candidate commits and iteratively pick the ones that jointly explain the fix. It introduces LinkRank, which scores candidates with a learning-to-rank model, then repeatedly selects the top commit, removes it, renormalizes the remaining scores, and stops by either a known count or an automatic threshold. On a new six-project dataset with large, imbalanced candidate pools, this formulation beats stronger neural and language-model baselines under both oracle and automatic stopping. A sympathetic reader cares because incomplete links break maintenance, debugging, and impact analysis; reframing the problem, not just scaling the model, is what closes the gap.

What carries the argument

LinkRank: treat each issue as a query over a candidate commit pool, score with a learning-to-rank model, then iteratively pick the highest-scoring commit, remove it, renormalize remaining scores, and stop via Known-K, absolute threshold (ABS), or relative threshold (REL).

What would settle it

Rebuild candidate pools without using ground-truth link timestamps (for example, from issue open time alone or a fixed repo-wide window), keep the same ranker and stopping rules, and check whether LinkRank still beats the strongest pairwise and LLM baselines on set-level F1 across the six projects.

Watch

Extended reading notes

Core claim

The authors claim that one-to-many issue–commit recovery is fundamentally an issue-centric set-recovery problem, not pairwise binary classification. When candidates are ranked jointly inside each issue’s pool and selected with an iterative pick–remove–renormalize loop, a lightweight ranker recovers complete commit sets more accurately than heavier methods built for one-to-one linking, under both Known-K and automatic Unknown-K stopping.

Load-bearing premise

Candidate pools are built from a time window centered on the earliest and latest ground-truth linked commits for that issue, so the search space is already focused on the real resolution period rather than built blindly from the full history.

Editorial extensions

If this is right

  • Traceability tools should output a variable-size commit set per issue, not a single best commit or independent pair labels.
  • Ranking quality inside an issue-specific pool matters more for multi-commit recovery than raw model size or pairwise classifier accuracy.
  • Automatic ABS/REL stopping can approach oracle Known-K performance without knowing how many commits belong to an issue.
  • New benchmarks should preserve full per-issue commit multiplicity and large realistic candidate pools, not only balanced pair labels.
  • Metadata such as time proximity and file overlap should be first-class features alongside text and embedding similarity.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same pick–remove–renormalize pattern may transfer to other one-to-many software links (pull request to issues, requirement to tests) where pairwise classifiers currently dominate.
  • If blind candidate generation erodes the gains, hybrid pipelines that first filter with cheap retrieval then rank would become the practical deployment path the authors gesture toward in future work.
  • Explainable set recovery—short rationales for why each selected commit belongs—could be the product surface developers actually trust, beyond F1 alone.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 0 minor

Summary. The paper reframes one-to-many issue–commit traceability as an issue-centric set-recovery problem rather than independent pairwise classification, and operationalizes this as LinkRank: LambdaMART ranking over 17 lexical/retrieval/semantic/metadata features within per-issue candidate pools, followed by an iterative pick–remove–renormalize selector with Known-K (oracle), ABS, and REL stopping rules. The authors construct a new dataset of 3,103 issues and 7,688 commits from six GitHub repositories (K=1..7), evaluate with 5-fold stratified CV, and report average Known-K F1 of 73.17% vs 48.39% for the strongest baseline (LinkAnchor), with Unknown-K F1 of 65.94% (ABS) and 62.19% (REL). The evaluation includes ranking metrics (MRR/NDCG@K/P@K), a feature ablation, threshold sensitivity analysis, a cost comparison, and Wilcoxon/Cliff's δ significance testing.

Significance. If the central comparison holds up under a blind candidate-pool construction, this is a useful contribution: the set-recovery reframing is clearly articulated, the dataset (3,103 issues / 7,688 commits, preserving raw artifacts and full per-issue commit sets, released via a public GitHub repository) fills a real gap left by tokenized releases such as EALink's, and the evaluation is unusually thorough for this literature — per-project tables with standard deviations, feature ablation showing metadata contributes the largest single gain (Table 7), threshold sensitivity plateaus (Fig. 4), a runtime comparison (Fig. 3), and effect-size reporting. The Unknown-K results with automatic stopping are the practically relevant numbers and remain strong. However, the headline 25-point F1 gap currently rests on a candidate-pool construction that uses ground-truth commit timestamps, so the magnitude of the improvement over baselines is not yet established in a deployment-realistic retrieval setting; the qualitative conclusion (issue-centric ranking beats pairwise classification) is likely robust, but this needs to be demonstrated rather than assumed.

major comments (1)
  1. Section 3.1.4, candidate pool definition P_i = {c : t_c in [t_min(C_i) - delta, t_max(C_i) + delta]}: the pool window is derived from the earliest and latest timestamps of the ground-truth linked commits, which is oracle information unavailable at recovery time. A deployed system knows the issue creation/close dates, not the linked commits' timestamps. This interacts directly with the paper's strongest feature signal: Table 7 shows metadata (f_time, f_dir relative to issue creation) contributes the largest ablation gain (49.79% -> 76.38% on PyTorch), and those temporal features are evaluated against negatives that are guaranteed to be temporally plausible because the pool is centered on the true resolution period. Part of the 73.17% vs 48.39% gap may therefore reflect pool geometry rather than the ranking formulation. The fix is within scope: re-run the main tables (4-6) with a blind RDS

Circularity Check

1 steps flagged · score 2.0 of 10

No circular derivation of the main claim; only a tautological coverage check in the RDS validity argument (pools already contain all ground-truth commits by construction).

  1. self definitional [§3.1.4 Candidate Pool Construction; §6.2 Limitations (candidate pool construction)]
    "P_i = {c ∈ C_repo | t_c ∈ [t_min(C_i) − δ, t_max(C_i) + δ]}, where t_min(C_i) and t_max(C_i) denote the earliest and latest timestamps among the ground-truth commits for issue i. ... We mitigate this by choosing the window based on the observed issue–commit time-gap distribution across all six projects and verifying that more than 98% of ground-truth commits fall within it."

    Because the RDS window edges are defined from the ground-truth set C_i itself, every c ∈ C_i satisfies t_c ∈ [t_min(C_i), t_max(C_i)] and therefore lies in the extended window for any δ ≥ 0. Coverage of ground-truth commits is 100% by construction. Presenting a “>98% fall within it” verification as independent evidence that the window is realistic is tautological: the check cannot fail and does not support the realism claim.

full rationale

LinkRank’s central claim is empirical, not a first-principles derivation: on held-out stratified folds, an issue-centric LambdaMART ranker with iterative pick–remove–renormalize selection beats pairwise/LLM baselines under Known-K and ABS/REL Unknown-K. Features (TF-IDF+SVD, BM25, SBERT, metadata), Optuna hyperparameter search, and τ/γ grid search are fit on development splits and evaluated on test folds—standard and not self-definitional. There is no uniqueness theorem, no load-bearing self-citation chain, and no renaming of a known closed-form result. The only circular step is peripheral: §3.1.4 defines each candidate pool from the ground-truth commit timestamps t_min(C_i) and t_max(C_i), so every true link is inside the pool by construction; §6.2 then cites “more than 98% of ground-truth commits fall within it” as mitigation evidence. That check is vacuous and does not validate realism. It does not force the reported F1 gaps (same pools are used for all methods), so it raises the score only to 2, not into the 6+ range reserved for claims that reduce to their inputs.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The claim rests on standard IR/LTR machinery plus several domain and evaluation choices: GitHub-derived links as ground truth, K in 1..7, RDS pools anchored on true link timestamps with δ=365 days, and dev-tuned stop thresholds. No new physical entities; free parameters are thresholds, window size, and boosted-tree hyperparameters.

free parameters (5)
  • ABS threshold τ = per-fold, often ~0.8–0.9 on sensitivity plots
    Grid-searched per fold on development F1 over [0.01,0.99]; directly controls Unknown-K set size and reported F1.
  • REL threshold γ = per-fold in [0.01,1]
    Same dev-set grid search; relative stop rule parameter for Unknown-K.
  • RDS temporal margin δ = 365 days
    Hand-set to 365 days to build candidate pools; changes pool difficulty and recall ceiling.
  • Max linked commits K_max = 7
    Issues with K>7 excluded by design; shapes dataset distribution and claim scope.
  • LambdaMART / LightGBM hyperparameters = fold-specific within stated ranges
    Trees, leaves, learning rate, child samples, feature/bagging fractions tuned with Optuna (50 trials) per fold; ranking quality depends on these fits.
assumptions (5)
  • domain assumption GitHub-visible issue–commit associations (via PRs/references) are sufficiently complete and correct ground truth for evaluation.
    Dataset construction and all F1 labels depend on this; missing manual links would bias both training and metrics (Sections 3.1, 6.2).
  • ad hoc to paper A candidate pool defined from the time span of true linked commits (±δ) is a realistic proxy for retrieval difficulty.
    Section 3.1.4 RDS definition; central to claiming ‘realistic’ evaluation while using label timestamps.
  • domain assumption Restricting to K≤7 preserves representative one-to-many behavior without material selection bias.
    Inclusion criteria in 3.1.3 and threats in 6.2; authors argue K>7 is <2% and noisy.
  • domain assumption Query-grouped LambdaMART optimizing ranking objectives is an appropriate model class for recovering commit sets from heterogeneous numeric features.
    Phase III justification; standard LTR assumption, not proved for this SE task beyond empirics.
  • ad hoc to paper Min–max normalization within the active pool plus iterative removal yields comparable scores for threshold stopping across issues.
    Algorithm 1 / Section 3.4 pick–remove–renormalize design choice underlying ABS/REL.
invented entities (1)
  • LinkRank iterative pick–remove–renormalize selector (Known-K / ABS / REL)
    purpose: Turn a per-issue ranked list into a variable-size predicted commit set for one-to-many recovery.
    Core procedural invention of the paper; evaluated empirically, not an external natural kind.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LinkRank: A Learning-to-Rank Framework for One-to-Many Issue-Commit Traceability." pith.science (2026). https://pith.science/paper/QT5VEFLB

@misc{pith2026260723610,
  author       = {Pith},
  title        = {Pith review of: LinkRank: A Learning-to-Rank Framework for One-to-Many Issue-Commit Traceability},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QT5VEFLB}},
  note         = {Machine review of arXiv:2607.23610}
}
read the original abstract

Recovering traceability links between issues and commits is important for software maintenance, debugging, impact analysis, and project understanding. However, most existing approaches assume a one-to-one relationship, where each issue is linked to a single commit. In practice, many issues are resolved through multiple commits, and ignoring this one-to-many nature can lead to incomplete traceability. This paper presents LinkRank, a learning-to-rank framework for recovering one-to-many issue--commit links. Unlike existing methods that mainly judge issue--commit pairs independently, LinkRank considers the set of candidate commits for an issue and identifies the commits that are most likely to contribute to its resolution. To support realistic evaluation, we construct a new dataset from six open-source GitHub repositories. LinkRank follows an iterative pick--remove--renormalize strategy: it selects the highest-ranked commit, removes it from the candidate pool, renormalizes the remaining scores, and repeats the process until the stopping criterion is met. We evaluate LinkRank under two settings: Known-K, where the true number of linked commits is available, and Unknown-K, where the model must infer when to stop selecting commits using ABS and REL stopping rules. Across six projects, LinkRank achieves an average Known-K F1 score of 74.54%, compared with 48.39% for the strongest baseline. In the Unknown-K setting, LinkRank achieves 68.84% F1 with ABS and 67.02% F1 with REL, outperforming the strongest baselines under both automatic stopping rules. Overall, the findings suggest that one-to-many issue--commit traceability is better addressed as an issue-centric ranking and iterative selection problem than as independent pairwise classification.

Figures

Figures reproduced from arXiv: 2607.23610 by the authors.

Figure 1
Figure 1. An example of one-to-many issue–commit traceabil￾ity, where a single issue (#35813) is resolved through a pull request (#35817) containing multiple focused commits. Each commit contributes to a different part of the issue resolution, such as test setup stabilization, broker startup synchronization, and teardown improvement. other commits that are also necessary to understand the full resolution of the issue. To addr… view at source ↗
Figure 2
Figure 2. Overview of the proposed LinkRank framework for one-to-many issue–commit recovery. Some recent datasets do contain issues linked to multiple commits. EALink [43], for instance, includes pair-level ex￾amples where some issues correspond to multiple commits, but its released form is preprocessed and tokenized rather than provided as raw artifacts. Consequently, information essential for realistic one-to-many recovery,… view at source ↗
Figure 3
Figure 3. compares the training and test-time costs of LinkRank with the baseline methods. The reported times are measured in minutes over the full evaluation pipeline. For methods that require a training phase, the left marker denotes [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Sensitivity of LinkRank to stopping thresholds 𝜏 (ABS) and 𝛾 (REL). Lines show mean F1 over 5-fold cross￾validation, bands show ±1 standard deviation, and dots indi￾cate development-tuned thresholds. The broad plateau shows that LinkRank is robust to threshold choice. …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 2 linked inside Pith

  1. [1]

    Linkanchor:Anautonomousllm-basedagentforissue-to-commitlink recovery

    Akhavan, A., Hosseinpour, A., Heydarnoori, A., Keshani, M., 2025. Linkanchor:Anautonomousllm-basedagentforissue-to-commitlink recovery. arXiv preprint arXiv:2508.12232

    arXiv 2025

  2. [2]

    Op- tuna: A next-generation hyperparameter optimization framework, in: Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pp

    Akiba, T., Sano, S., Yanase, T., Ohta, T., Koyama, M., 2019. Op- tuna: A next-generation hyperparameter optimization framework, in: Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pp. 2623–2631

    2019

  3. [3]

    Pi-link: A ground-truth dataset of links between pull- requests and issues in github

    Alshara, Z., Shatnawi, A., Eyal-Salman, H., Seriai, A.D., Shatnawi, M., 2022. Pi-link: A ground-truth dataset of links between pull- requests and issues in github. IEEE Access 11, 697–710

    2022

  4. [4]

    A literature review of auto- matictraceabilitylinksrecoveryforsoftwarechangeimpactanalysis, in: Proceedings of the 28th International Conference on Program Comprehension, pp

    Aung, T.W.W., Huo, H., Sui, Y., 2020. A literature review of auto- matictraceabilitylinksrecoveryforsoftwarechangeimpactanalysis, in: Proceedings of the 28th International Conference on Program Comprehension, pp. 14–24

    2020

  5. [5]

    Bachmann, A., Bernstein, A., 2009. Software process data quality and characteristics: a historical view on open and closed source projects,in:ProceedingsofthejointinternationalandannualERCIM workshopsonPrinciplesofsoftwareevolution(IWPSE)andsoftware evolution (Evol) workshops, pp. 119–128

    2009

  6. [6]

    Bachmann, A., Bird, C., Rahman, F., Devanbu, P., Bernstein, A.,

  7. [7]

    Bird, C., Bachmann, A., Rahman, F., Bernstein, A., 2010. Linkster: enablingefficientmanualinspectionandannotationofmineddata,in: Proceedings of the eighteenth ACM SIGSOFT international sympo- sium on Foundations of software engineering, pp. 369–370

    2010

  8. [8]

    Howdo centralized and distributed version control systems impact software changes?, in: Proceedings of the 36th international conference on Software Engineering, pp

    Brindescu,C.,Codoban,M.,Shmarkatiuk,S.,Dig,D.,2014. Howdo centralized and distributed version control systems impact software changes?, in: Proceedings of the 36th international conference on Software Engineering, pp. 322–333

    2014

Show all 49 references
  1. [9]

    From ranknet to lambdarank to lambdamart: An overview

    Burges, C.J., 2010. From ranknet to lambdarank to lambdamart: An overview. Learning 11, 23–581

    2010

  2. [10]

    Dominance statistics: Ordinal analyses to answer ordinal questions

    Cliff, N., 1993. Dominance statistics: Ordinal analyses to answer ordinal questions. Psychological bulletin 114, 494

    1993

  3. [11]

    Cortés-Coy, L.F., Linares-Vásquez, M., Aponte, J., Poshyvanyk, D.,

  4. [12]

    Ratelimitsfortherestapi

    GitHub,2022. Ratelimitsfortherestapi. URL:https://docs.github. com/en/rest/using-the-rest-api/rate-limits-for-the-rest-api? apiVersion=2022-11-28. accessed: February 7, 2025. Abhishek Kumar et al. Page 14 of 15

    2022

  5. [13]

    Backtothe basics: Rethinking issue-commit linking with llm-assisted retrieval

    Huang, H., Widyasari, R., Zhang, T., Irsan, I.C., Shi, J., Ang, H.W., Liauw,F.,Ouh,E.L.,Shar,L.K.,Kang,H.J.,etal.,2025. Backtothe basics: Rethinking issue-commit linking with llm-assisted retrieval. arXiv preprint arXiv:2507.09199

    2025

  6. [14]

    Link- former: Automatic contextualised link recovery of software artifacts in both project-based and transfer learning settings

    Izadi, M., Mazrae, P.R., Mens, T., van Deursen, A., 2022. Link- former: Automatic contextualised link recovery of software artifacts in both project-based and transfer learning settings. arXiv preprint arXiv:2211.00381

    2022 arXiv

  7. [15]

    Lightgbm: A highly efficient gradient boosting decision tree

    Ke,G.,Meng,Q.,Finley,T.,Wang,T.,Chen,W.,Ma,W.,Ye,Q.,Liu, T.Y., 2017. Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems 30

    2017

  8. [16]

    Linkrank:Learning-to-rankforone-to-manyissue– commit traceability recovery.https://github.com/merealone2516/ LinkRank

    Kumar,A.,2026. Linkrank:Learning-to-rankforone-to-manyissue– commit traceability recovery.https://github.com/merealone2516/ LinkRank. GitHub repository

    2026

  9. [17]

    Btlink: automatic link recovery between issues and commits based on pre-trained bert model

    Lan, J., Gong, L., Zhang, J., Zhang, H., 2023. Btlink: automatic link recovery between issues and commits based on pre-trained bert model. Empirical Software Engineering 28, 1–55

    2023

  10. [18]

    Rclinker:Automatedlinkingofissuereportsandcommitsleveraging richcontextualinformation,in:2015IEEE23rdInternationalConfer- ence on Program Comprehension, IEEE

    Le, T.D.B., Linares-Vásquez, M., Lo, D., Poshyvanyk, D., 2015. Rclinker:Automatedlinkingofissuereportsandcommitsleveraging richcontextualinformation,in:2015IEEE23rdInternationalConfer- ence on Program Comprehension, IEEE. pp. 36–47

    2015

  11. [19]

    Trace- ability transformed: Generating more accurate links with pre-trained bert models, in: 2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE), IEEE

    Lin, J., Liu, Y., Zeng, Q., Jiang, M., Cleland-Huang, J., 2021. Trace- ability transformed: Generating more accurate links with pre-trained bert models, in: 2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE), IEEE. pp. 324–335

    2021

  12. [20]

    Linares-Vásquez, M., Cortés-Coy, L.F., Aponte, J., Poshyvanyk, D.,

  13. [21]

    O’Reilly Media, Inc

    Loeliger, J., McCullough, M., 2012. Version Control with Git: Pow- erful tools and techniques for collaborative software development. " O’Reilly Media, Inc."

    2012

  14. [22]

    Mahmoud,A.,Niu,N.,Xu,S.,2012.Asemanticrelatednessapproach for traceability link recovery, in: 2012 20th IEEE international con- ference on program comprehension (ICPC), IEEE. pp. 183–192

    2012

  15. [23]

    Automated recovery ofissue-commitlinksleveragingbothtextualandnon-textualdata,in: 2021 IEEE International Conference on Software Maintenance and Evolution (ICSME), IEEE

    Mazrae, P.R., Izadi, M., Heydarnoori, A., 2021. Automated recovery ofissue-commitlinksleveragingbothtextualandnon-textualdata,in: 2021 IEEE International Conference on Software Maintenance and Evolution (ICSME), IEEE. pp. 263–273

    2021

  16. [24]

    Using traceability to sup- port model-based regression testing, in: Proceedings of the 22nd IEEE/ACM International Conference on Automated Software Engi- neering, pp

    Naslavsky, L., Richardson, D.J., 2007. Using traceability to sup- port model-based regression testing, in: Proceedings of the 22nd IEEE/ACM International Conference on Automated Software Engi- neering, pp. 567–570

    2007

  17. [25]

    Nguyen, A.T., Nguyen, T.T., Nguyen, H.A., Nguyen, T.N., 2012. Multi-layered approach for recovering links between bug reports and fixes, in: Proceedings of the ACM SIGSOFT 20th International Symposium on the Foundations of Software Engineering, pp. 1–11

    2012

  18. [26]

    Hermes: Using commit-issue linking to detect vulnerability-fixing commits, in: 2022 IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), IEEE

    Nguyen-Truong, G., Kang, H.J., Lo, D., Sharma, A., Santosa, A.E., Sharma, A., Ang, M.Y., 2022. Hermes: Using commit-issue linking to detect vulnerability-fixing commits, in: 2022 IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), IEEE. pp. 51–62

    2022

  19. [27]

    Successfuldeploymentofrequirementstraceabil- ity in a commercial engineering organization

    Panis,M.C.,2010. Successfuldeploymentofrequirementstraceabil- ity in a commercial engineering organization... really, in: 2010 18th IEEEInternationalRequirementsEngineeringConference,IEEE.pp. 303–307

    2010

  20. [28]

    Traceability in the wild: automatically augmenting incomplete trace links, in: Proceedings of the 40th International Conference on Soft- ware Engineering, pp

    Rath, M., Rendall, J., Guo, J.L., Cleland-Huang, J., Mäder, P., 2018. Traceability in the wild: automatically augmenting incomplete trace links, in: Proceedings of the 40th International Conference on Soft- ware Engineering, pp. 834–845

    2018

  21. [29]

    3982–3992

    Reimers,N.,Gurevych,I.,2019.Sentence-bert:Sentenceembeddings using siamese bert-networks, in: Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th in- ternationaljointconferenceonnaturallanguageprocessing(EMNLP- IJCNLP), pp. 3982–3992

    2019

  22. [30]

    Developing Safety-Critical Software: A Practical Guide for Aviation Software and DO-178C Compliance

    Rierson, L., 2013. Developing Safety-Critical Software: A Practical Guide for Aviation Software and DO-178C Compliance. CRC Press

    2013

  23. [31]

    The probabilistic relevance framework: Bm25 and beyond

    Robertson, S., Zaragoza, H., et al., 2009. The probabilistic relevance framework: Bm25 and beyond. Foundations and Trends®in Infor- mation Retrieval 3, 333–389

    2009

  24. [32]

    Deeplink: Recovering issue-commit links based on deep learning

    Ruan, H., Chen, B., Peng, X., Zhao, W., 2019. Deeplink: Recovering issue-commit links based on deep learning. Journal of Systems and Software 158, 110406

    2019

  25. [33]

    Schermann, G., Brandtner, M., Panichella, S., Leitner, P., Gall, H.,

  26. [34]

    Theproductbacklog,in:2019 IEEE/ACM 41st International Conference on Software Engineering (ICSE), IEEE

    Sedano,T.,Ralph,P.,Péraire,C.,2019. Theproductbacklog,in:2019 IEEE/ACM 41st International Conference on Software Engineering (ICSE), IEEE. pp. 200–211

    2019

  27. [35]

    Bugzilla, itracker, and other bug trackers

    Serrano, N., Ciordia, I., 2005. Bugzilla, itracker, and other bug trackers. IEEE software 22, 11–13

    2005

  28. [36]

    Cross-validatorychoiceandassessmentofstatistical predictions

    Stone,M.,1974. Cross-validatorychoiceandassessmentofstatistical predictions. Journaloftheroyalstatisticalsociety:SeriesB(Method- ological) 36, 111–133

    1974

  29. [37]

    Discoveringlonersandphantomsincommitandissuedata,in: 2015 IEEE 23rd International Conference on Program Comprehen- sion, IEEE. pp. 4–14

    2015

  30. [38]

    Frlink:Improvingtherecoveryof missing issue-commit links by revisiting file relevance

    Sun,Y.,Wang,Q.,Yang,Y.,2017b. Frlink:Improvingtherecoveryof missing issue-commit links by revisiting file relevance. Information and Software Technology 84, 33–47

  31. [39]

    Mplinker:Multi- template prompt-tuning with adversarial training for issue-commit link recovery

    Wang,B.,Deng,Y.,Luo,R.,Liang,P.,Bi,T.,2025. Mplinker:Multi- template prompt-tuning with adversarial training for issue-commit link recovery. Journal of Systems and Software , 112351

    2025

  32. [40]

    Wilcoxonsigned-ranktest

    Woolson,R.F.,2007. Wilcoxonsigned-ranktest. Wileyencyclopedia of clinical trials , 1–3

    2007

  33. [41]

    Improving miss- ing issue-commit link recovery using positive and unlabeled data, in: 2017 32nd IEEE/ACM International Conference on Automated Software Engineering (ASE), IEEE

    Sun, Y., Chen, C., Wang, Q., Boehm, B., 2017a. Improving miss- ing issue-commit link recovery using positive and unlabeled data, in: 2017 32nd IEEE/ACM International Conference on Automated Software Engineering (ASE), IEEE. pp. 147–152

    2017

  34. [42]

    Xie, R., Chen, L., Ye, W., Li, Z., Hu, T., Du, D., Zhang, S., 2019. Deeplink:Acodeknowledgegraphbaseddeeplearningapproachfor issue-commitlinkrecovery,in:2019IEEE26thInternationalConfer- ence on Software Analysis, Evolution and Reengineering (SANER), IEEE. pp. 434–444

    2019

  35. [43]

    Zhang, C., Wang, Y., Wei, Z., Xu, Y., Wang, J., Li, H., Ji, R.,

  36. [44]

    Versioncontrolsystem: A review

    Zolkifli,N.N.,Ngah,A.,Deraman,A.,2018. Versioncontrolsystem: A review. Procedia Computer Science 135, 408–415. Abhishek Kumar et al. Page 15 of 15

    2018

  37. [45]

    Relink: recovering links between bugs and changes, in: Proceedings of the 19th ACM SIGSOFTsymposiumandthe13thEuropeanconferenceonFounda- tions of software engineering, pp

    Wu, R., Zhang, H., Kim, S., Cheung, S.C., 2011. Relink: recovering links between bugs and changes, in: Proceedings of the 19th ACM SIGSOFTsymposiumandthe13thEuropeanconferenceonFounda- tions of software engineering, pp. 15–25

    2011

  38. [2010]

    The missing links: bugs and bug-fix commits, in: Proceed- ings of the eighteenth ACM SIGSOFT international symposium on Foundations of software engineering, pp. 97–106

  39. [2014]

    On automatically generating commit messages via summa- rization of source code changes, in: 2014 IEEE 14th International Working Conference on Source Code Analysis and Manipulation, IEEE. pp. 275–284

    2014

  40. [2015]

    Changescribe: A tool for automatically generating commit messages, in: 2015 IEEE/ACM 37th IEEE International Conference on Software Engineering, IEEE. pp. 709–712

    2015

  41. [2023]

    Ealink: An efficient and accurate pre-trained framework for issue-commit link recovery, in: 2023 38th IEEE/ACM International Conference on Automated Software Engineering (ASE), IEEE. pp. 217–229

    2023

Pith tools

Reviewed July 30, 2026 · model on record in the stance chip above.