Intrusion detection method combining stacked autoencoder and FSVM

Large-scale, high-dimensional sample data have a strong impact on the learning speed and learning effect of the l, so an intrusion detection method combining stacked auto-encoder and fast support vector machine (SAE-FSVM) was come up for discussion. A dimensionality reduction model based on stacked auto-encoder was built, through which the network data was extracted with features to obtain high-quality feature data. The clustering algorithm was used to cluster different types of data in the samples separately. The similarity of clusters between different types of samples was calculated to obtain similar clusters between different types of samples, and the sample data in the similar clusters were used as the input of the support vector machine model, so as to reduce the scale of the input samples and improved the learning speed of the model. Through simulation experiments, it is verified that this method can effectively improve the learning speed and learning effect of the support vector machine model for large-scale high-dimensional data.