Assisting the appraisal of e-mail records with automatic classification
Bibliographic record
Abstract
Purpose This paper aims to investigate how automatic classification can assist employees and records managers with the appraisal of e-mails as records of value for the organization. Design/methodology/approach The study performed a qualitative analysis of the appraisal behaviours of eight records management experts to train a series of support vector machine classifiers to replicate the decision process for identifying e-mails of business value. Automatic classification experiments were performed on a corpus of 846 e-mails from two of these experts’ mailboxes. Findings Despite the highly contextual nature of record value, these experiments show that classifiers have a high degree of accuracy. Unlike existing manual practices in corporate e-mail archiving, machine classification models are not highly dependent on features such as the identity of the sender and receiver or on threading, forwarding or importance flags. Rather, the dominant discriminating features are textual features from the e-mail body and subject field. Research limitations/implications The need to automatically classify corporate e-mails is growing in importance, as e-mail remains one of the prevalent recordkeeping challenges. Practical implications Automated methods for identifying e-mail records promise to be of significant benefit to organizations that need to appraise e-mail for long-term preservation and access on demand. Social implications The research adopts an innovative approach to assist employees and records managers with the appraisal of digital records. By doing so, the research fosters new insights on the adoption of technological strategies to automate recordkeeping tasks, an important research gap. Originality/value Our experiment show that a SVM classifier can be trained to replicate an expert's decision process for identifying e-mails of business value with a reasonably high degree of accuracy. In principle, such a classifier could be integrated into a corporate Electronic Document and Records Management System (EDRMS) to improve the quality of e-mail records appraisal.
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 | 0.016 | 0.110 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.005 | 0.005 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.004 |
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".