On Some Feature Selection Strategies for Spam Filter Design
Bibliographic record
Abstract
Feature selection is an important research problem in different statistical learning problems including text categorization applications such as spam email classification. In designing spam filters, we often represent the email by vector space model (VSM), i.e., every email is considered as a vector of word terms. Since there are many different terms in the email, and not all classifiers can handle such a high dimension, only the most powerful discriminatory terms should be used. Another reason is that some of these features may not be influential and might carry redundant information which may confuse the classifier. Thus, feature selection, and hence dimensionality reduction, is a crucial step to get the best out of the constructed features. There are many feature selection strategies that can be applied to produce the resulting feature set. In this paper, we investigate the use of hill climbing, simulated annealing, and threshold accepting optimization techniques as feature selection algorithms. We also compare the performance of the above three techniques with the linear discriminate analysis. Our experiment results show that all these techniques can be used not only to reduce the dimensions of the e-mail, but also improve the performance of the classification filter. Among all the strategies, simulated annealing has the best performance which reaches a classification accuracy of 95.5%
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 | 0.002 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".