Scholarlink

Intrusion detection systems (IDS) provide effective identification of malicious activity occurring on a network. In this paper we discuss the process of creating an IDS using machine learning techniques. We first acquired and cleaned the datasets for any redundant data and any values holding 0. We used data visualization to obtain better insights on what the dataset is doing, and which variables affect each other. Testing was performed using machine learning algorithms specifically Logistic Regression, Decision Tree Classifier, Naïve Bayes, Random Forest Classifier and K-Nearest Neighbors for performance measuring. We then tuned the model for better metric performance. Finally, we developed the IDS based on one of the models and generated the alerts to predict if there is an attack. Our goal is for the IDS to accurately predict the number of attacks.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using Machine Learning Techniques to Detect Network Intrusions

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jawaan Clifton,
  • Cheryl Hinds

摘要

Intrusion detection systems (IDS) provide effective identification of malicious activity occurring on a network. In this paper we discuss the process of creating an IDS using machine learning techniques. We first acquired and cleaned the datasets for any redundant data and any values holding 0. We used data visualization to obtain better insights on what the dataset is doing, and which variables affect each other. Testing was performed using machine learning algorithms specifically Logistic Regression, Decision Tree Classifier, Naïve Bayes, Random Forest Classifier and K-Nearest Neighbors for performance measuring. We then tuned the model for better metric performance. Finally, we developed the IDS based on one of the models and generated the alerts to predict if there is an attack. Our goal is for the IDS to accurately predict the number of attacks.