UNSUPERVISED EXCEPTIONAL ACCESS DETECTION METHOD AND APPARATUS BASED ON ONE-HOT ENCODING MECHANISM

摘要

Disclosed in the present application are an unsupervised exceptional access detection method and apparatus based on a one-hot encoding mechanism. According to the method, an exceptional URL can be accurately detected by using a bigram model, a one-hot encoding mechanism, and a deep autoencoder network while the characteristics of the exceptional URL are unknown, thereby avoiding exceptional access and avoiding harm caused by malicious access. In addition, the problem that it is difficult to accurately detect the exceptional URL by using a fixed rule can be resolved. The characteristics of high detection accuracy and good robustness are provided. The present application can be widely applied to the technical field of next-generation Internet security such as exceptional access detection and exceptional flow detection. By means of unsupervised learning, the exceptional URL can be accurately recognized while the characteristics of the exceptional URL are ambiguous and there are few exceptional samples. Moreover, the training stage of the deep autoencoder network can be completed offline, and after the deep autoencoder network is established, a high detection speed is achieved, thereby greatly improving the efficiency of exceptional access detection.