Research on Industrial Control Network Security Based on Automatic Machine Learning
Authors:
- 25 November 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 6, 149-153
Abstract
Aiming at the problem of complex parameter configuration and slow effective model construction in industrial control system attack identification, starting from the study of the whole process of machine learning, an automatic machine learning (AUTOML) attack behavior identification method is proposed, which takes KDD99 industrial control network data as the detection object. The tree structure based pipeline optimization (TPOT) method, which automatically selects machine learning algorithm and parameter optimization through genetic programming to construct the optimal model pipeline, is used to predict the attack behavior of industrial control network. The empirical study shows that, compared with support vector machine, Naive Bayes, KNN and Logistic Regression, Compared with this method, traditional machine learning methods such as Adaboost reduce the human participation and save a lot of manpower, material resources and time. At the same time, the average accuracy of 10 tests reaches 92.62 % and the average F3_score index reaches 92.17%.Keywords
This publication has 6 references indexed in Scilit:
- Intelligent intrusion detection system using multilayer perceptron optimised by genetic algorithmInternational Journal of Computational Intelligence Studies, 2020
- Anomaly-based Intrusion Detection in Industrial Data with SVM and Random Forests2019
- Machine learning meets continuous flow chemistry: Automated optimization towards the Pareto front of multiple objectivesChemical Engineering Journal, 2018
- Multi-objective optimization of base classifiers in StackingC by NSGA-II for intrusion detection2017
- Evaluation of genetic programming-based models for simulating bead dimensions in wire and arc additive manufacturingJournal of Intelligent Manufacturing, 2016
- Evaluation of a Tree-based Pipeline Optimization Tool for Automating Data Science2016