Question-And-Answer Community Mining in Software Project Management – A Deep Learning Approach
Bibliographic record
Abstract
Software project management (SPM) is one of the most dominant fields in Software Engineering (SE). During recent years, excessive growth in data science has brought a new research opportunity for supporting project managers, referred to as SPM Analytics. The majority of the field efforts are concerned with using projects' data in the estimation problems and mining the general public data in Requirement Engineering applications. However, in more general SE applications, Question and Answer (QA) communities such as StackOverflow have been known as a rich data source. While most studies in SPM analytics use traditional Machine Learning (ML) methods, in this thesis, a method named DeepQA-Miner based on Deep Neural Networks (DNNs) is proposed to mine SPM QA communities. Project Management StackExchange (PMSE), a well-known community for project managers, is targeted. It provides project managers with the opportunity to share their questions, making it a great candidate for characterizing practitioners' needs. The DeepQA-Miner method would pre-process the data and feed it into a multi-input multi-head network. The network receives different data parts separately, embeds the text internally, extracts the essential patterns, and classifies it for multi-purposes, leveraging a single shared knowledge base. More than 5000 questions at PMSE are accessed, classified through four different perspectives, and analyzed by their tone to formulate SPM practitioners' needs. The DeepQA-Miner's performance is compared with four baseline methods. Overall, DeepQA-Miner outperforms the other classifiers. Even though two of the traditional methods achieved slightly higher accuracy in one of the binary classification tasks, there is a remarkable improvement by the DeepQA-Miner in multi-class tasks. Furthermore, the findings provide potential directions for further research and development. As an application, the findings are compared with SPM education status quo resulting from SPM-related courses in Canada's top 10 universities. A set of considerations for reducing the existing gap between the industry needs and courses' agenda is proposed. As a contribution to Open Science, all data parts are being made publicly available: https://github.com/alirzahmadi/DeepQA-Miner
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 | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| 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".