Elite-driven support vector machines for classification
Bibliographic record
Abstract
In the field of Support Vector Machines (SVM), the traditional approach to classifier construction relies heavily on a set of observations known as support vectors, determined by the choice of loss functions. Each loss function results in a specific decision boundary and identifies a unique set of support vectors, leading to varied classification performances. This thesis proposes a novel SVM methodology that gives additional weight to a curated collection of elite observations that play crucial roles in constructing SVM decision boundaries under various loss functions. These elite observations are identified through their recurring presence as support vectors in different SVM configurations using diverse loss functions. We develop new loss functions to emphasize the importance of these elite observations during the training of our SVM classifiers. The loss functions for the Elite-Driven Support Vector Machine (EDSVM) are designed to be classification-calibrated, ensuring theoretical soundness while enhancing the model's focus on these elite observations. Rigorous theoretical results are provided, and a comprehensive numerical data analysis is conducted to evaluate the EDSVM's performance across various datasets. The novel SVM models developed in this research demonstrate superior performance compared to conventional models studied in this thesis. This is evidenced through both simulation studies and real data analyses, applicable to both linear and non-linear classification tasks.
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 | 0.003 | 0.010 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.003 |
| 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".