Machine Learning-Based Alert Correlation for Enhanced Cybersecurity: A Multi-Classification Approach
Bibliographic record
Abstract
Modern Security Operations Centers (SOCs) struggle with an overwhelming influx of security alerts from diverse sources, leading to alert fatigue and delayed incident response times. Conventional Security Information and Event Management (SIEM) systems typically process alerts in isolation, missing complex attack patterns that unfold across multiple detection points. This research introduces a comprehensive machine learning framework for intelligent alert correlation using three classification approaches: binary correlation detection, correlation type classification, and incident classification. We evaluated eight machine learning algorithms—Logistic Regression, Random Forest, XGBoost, LightGBM, Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), CatBoost, and Support Vector Machines (SVM)—using 50,000 realistic security alerts generated from 2.30 million network flows in the CIC-IDS-2017 dataset. Our multi-classification framework incorporates temporal relationships, network topology, attack progression patterns, and severity indicators to enhance correlation precision. The experimental results reveal remarkable performance across all classification tasks. CatBoost achieved the highest overall performance with 99.27% average accuracy across all tasks, closely followed by Random Forest (99.17%) and LightGBM (99.06%). For multiclass correlation type classification, six out of eight models achieved perfect 100% accuracy, with only XGBoost and SVM showing decreases to 96.30%. Binary correlation detection demonstrated exceptional performance with three models (Random Forest, LightGBM, CatBoost) achieving 99.96% accuracy. The proposed framework demonstrates exceptional capability in reducing alert noise while maintaining optimal detection rates, significantly enhancing SOC operational efficiency and enabling more targeted threat response strategies.
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 | 0.004 | 0.006 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.005 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 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".