Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jul 23, 2025

Automated Detection and Analysis of Exocytosis

Automated Detection and Analysis of Exocytosis

Published on: September 11, 2021

SELID: Selective Event Labeling for Intrusion Detection Datasets.

Woohyuk Jang¹, Hyunmin Kim¹, Hyungbin Seo¹

¹Department of Computer Science, Kookmin University, 77, Jeongneung-ro, Seongbuk-gu, Seoul 02707, Republic of Korea.

Sensors (Basel, Switzerland)

|July 14, 2023

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Selective histone deacetylase inhibition as a therapeutically relevant strategy in H3K27M-glioma.

Neuro-oncology·2026

Same author

AI-driven telerehabilitation for older adults with mild cognitive impairment: a randomized controlled trial.

Frontiers in neurology·2026

Same author

RNA exosome-mediated RNA surveillance governs developmental timing in the human cerebellum.

bioRxiv : the preprint server for biology·2026

Same author

Considering the missing science of retraining and maintenance in medical artificial intelligence, using ophthalmology as an exemplar.

NPJ digital medicine·2026

Same author

Structural Comparison of the Human G93A Mutant SOD1 to the Wild-type SOD1 Filaments.

Journal of molecular biology·2026

Same author

Mitigating semantic label divergence in federated learning: Obfuscated encoding and alert filtering for security monitoring.

PloS one·2025

Security operations centers face alert fatigue due to overwhelming event volumes. This study introduces efficient data labeling using unsupervised clustering, reducing labeling needs to just 2% without impacting machine learning model performance.

Area of Science:

Cybersecurity
Machine Learning
Data Science

Background:

Security operations centers (SOCs) are overwhelmed by a high volume of security events, leading to alert fatigue among human analysts.
Machine learning (ML) offers a solution by automating the distinction between true and false security alerts.
Effective ML model training requires accurately labeled datasets, but manual labeling is resource-intensive.

Purpose of the Study:

To propose a novel selective sampling scheme for efficient data labeling in cybersecurity.
To reduce the human resources required for labeling large datasets for ML-based security event analysis.

Main Methods:

Transforming security event byte sequences into fixed-size vectors using content-defined chunking and feature hashing.
Applying unsupervised clustering algorithms to these vectors.

Keywords:

alert fatigue cyber security intrusion detection security operations center unsupervised learning

More Related Videos

Design and Analysis for Fall Detection System Simplification

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Jul 23, 2025

Automated Detection and Analysis of Exocytosis

Automated Detection and Analysis of Exocytosis

Published on: September 11, 2021

Design and Analysis for Fall Detection System Simplification

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Selecting a small subset of samples from each cluster for manual labeling.

Main Results:

The proposed selective sampling scheme significantly reduces the data labeling effort, requiring only 2% of the total data.
Experimental results show no degradation in the F1-score of the machine learning model despite the reduced labeling.
Validation performed on both a private SOC dataset and a public internet dataset.

Conclusions:

The developed selective sampling scheme offers an efficient method for data labeling in cybersecurity.
This approach effectively mitigates alert fatigue by enabling faster and more accurate ML model training.
The technique is validated for practical application in real-world security operations.