论文部分内容阅读
网络入侵的早期特征是影响网络入侵早期检测效果的关键.针对网络入侵早期特征选择问题,提出一种结合频率筛选的遗传算法,该算法以SOM神经网络作为评价模型,通过多次运行遗传算法改善其优化结果的稳定性,根据对最优解中特征出现的频率进一步筛选,得到一组优化的早期特征.对入侵早期特征集进行特征选择实验,将39维早期特征优化至29维.实验结果表明,使用优化特征组合不仅有效缩减了入侵检测建模时间,而且使入侵检测系统获得更高的检测率.
The early feature of network intrusion is the key to affect the early detection of network intrusion.Aiming at the early feature selection of network intrusion, a genetic algorithm combined with frequency filtering is proposed, which uses SOM neural network as evaluation model and improves genetic algorithm through multiple runs The stability of the optimization results was further screened according to the frequencies of the features appearing in the optimal solution to obtain a set of optimized early features.The feature selection experiments of early invasion feature sets were conducted to optimize the early features of 39 dimensions to 29 dimensions.Experimental results It shows that the use of optimized feature combination not only effectively reduces the intrusion detection modeling time, but also enables the intrusion detection system to achieve a higher detection rate.