Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces MR-Size, an explainable effort estimator that predicts T-shirt sizes and interpolated day estimates for GitLab merge requests by computing a composite complexity score based on code diffs, file weights, contributor dynamics, and semantic signals. MR-Size achieves a Pearson correlation of 0.79 and a mean absolute error of 2.34 days across 150 merge requests, matching LOC baselines while providing per-file explanations. The method, datasets, evaluation plan, and reproducibility artifacts are described, with a benchmarking protocol comparing MR-Size against LOC baselines, COCOMO-style models, and learned regressors.
Ditch the guesswork in agile planning: MR-Size automatically estimates effort from merge requests with per-file explanations, matching LOC baselines.
Abstract - Automated effort estimation that maps engineering work to agile sizing units (e.g., T-shirt sizes) would greatly streamline planning and release forecasting. We propose MR-Size, an explainable, repository-driven estimator that computes a composite complexity score for GitLab merge requests using code diffs, per-file weights, contributor dynamics, and semantic contextual signals (keywords and text embeddings). The estimator produces T-shirt sizes and interpolated day estimates while providing per-file explanations to maintain interpretability. Across 150 merge requests, MR-Size achieved a Pearson correlation of 0.79 and a mean absolute error of 2.34 days, matching LOC baselines while offering per-file explanations. This paper describes the method, datasets, planned evaluation, and reproducibility artifacts. We outline an empirical protocol comparing MR-Size against LOC baselines, COCOMO-style models, and learned regressors (e.g., XGBoost). The contributions are (1) a reproducible MR→T-shirt pipeline, (2) an explainability-first complexity formulation, and (3) a benchmarking plan across open-source and industrial repositories. Keywords: Effort Estimation, Merge Request, T-Shirt Sizing, Repository Mining, Explainability