XGBoost-RF的物联网入侵检测模型

Intrusion Detection Model of Internet of Things Based on XGBoost-RF

针对物联网入侵检测中检测数据不平衡导致的分类不准确的问题,提出了一种基于极端梯度提升树和随机森林相结合的物联网入侵检测模型.首先,针对物联网应用环境中产生的大量数据,对数据进行数据归一化处理.然后,利用XGBoost算法对其中的特征进行重要性评分,选择最优特征.最后,结合改进的随机森林算法,解决因数据不平衡导致的分类不准确的问题.仿真试验表明所提模型能有效的进行数据最优特征选择及合理地检测分类,同RF算法、SVM算法、Tree-SVM模型和RF-GDBT模型相比,所提模型的检测准确率有效改善.

In view of the problem of inaccurate classification caused by unbalanced detection data in the internet of things intrusion detection, an internet of things intrusion detection model based on the combination of extreme gradient boosting and the random forest is proposed.Firstly, due to the large amount of data generated in the application environment of the Internet of things, data normalization processing is carried out on the data.Then, the XGBoost algorithm is used to score the importance of the features and select the best features.Finally, combined with the improved random forest algorithm, the problem of inaccurate classification caused by data imbalance is solved.The simulation test shows that the optimal features of data can be effectively selected and reasonably classified by the proposed model.Compared with the RF algorithm, SVM algorithm, tree-SVM model and RF-GDBT model, the detection accuracy of the proposed model is effectively improved.