Rethinking Technological Investment and Cost-Benefit: A Software Requirements Dependency Extraction Case Study
Bibliographic record
Abstract
Machine Learning (ML) is widely used for different purposes within Software Engineering. It can substantially improve the efficiency and effectiveness of organizations. While various methods and techniques exist, all of them have strengths and weaknesses under varying scenarios and contexts. Thus far, the selection and implementation of ML techniques rely almost exclusively on accuracy criteria. This narrow perspective ignores crucial considerations of anticipated costs of the ML activities versus the projected benefits gained from applying the results. Thus, in this study we introduce a return-on-investment (ROI) perspective to evaluate ML techniques in Software Engineering, offering a novel lens to assess their true value beyond traditional benchmarks. We present findings for an approach that addresses this gap by enhancing the accuracy criterion with return on investment (ROI) considerations. Specifically, we extract dependencies from textual descriptions of software requirements and analyze the performance of two state-of-the-art ML techniques: Random Forest and Bidirectional Encoder Representations from Transformers (BERT), a encoder only Large Language Model. Drawing upon two publicly available data sets, we compare decision-making based on 1) exclusively on accuracy and 2) on ROI analysis to provide decision support for the selection and usage of ML classification methods. As such, our results showed that, 1) chasing model accuracy improvisation through increased annotated data does not generate expected returns in traditional ML methods. 2) For complex ML algorithms, the need for larger annotated dataset investment cost is justified by the higher returns, however, the trade-offs between accuracy and ROI become evident.
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 | 0.010 | 0.037 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| 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".