Pushing Feelings: Emotion and Sentiment in Software Commit Messages
Bibliographic record
Abstract
Software repositories are the primary source of information for software artifacts and contain a large amount of unstructured data, including commits, issue reports, and code comments. Mining information for different tasks, such as sentiment and emotion analysis, has been studied on several artifacts. While significant progress has been made in sentiment analysis on various artifacts, the study of emotion analysis remains under-researched due to a lack of resources, such as relevant and labeled datasets. This study presents a manually annotated dataset comprising 12,005 commit messages from the open-source Apache Tomcat project, exploring emotion and sentiment analysis through the application of natural language processing and machine learning techniques. We studied traditional models (SVM and random forest), deep learning models (bidirectional-LSTM), and pre-trained language models (LMs), as well as their ensemble, to evaluate and compare their performance on the curated dataset. Our findings suggest that software engineering domain-specific pretrained LMs consistently outperform traditional and deep learning models. The ensemble of pretrained LMs performs better on sentiment analysis. Additionally, identifying sentiment from commit messages outperforms emotion analysis tasks. We have made model implementations and the curated dataset available.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".