Research On Neural Network Algorithm Of Cloud Service Anomaly Detection

The continuous emergence and rapid application of these new technologies such as cloud computing,big data,artificial intelligence,etc.not only promoted the informatization process in China,but also made Internet technology gradually penetrate into all aspects of human social production and daily life.The rich resources brought by information technology enrich and facilitate human daily life,but while humans enjoy the convenience brought by information technology,due to the free and open nature of computer networks and the imperfect management system,they have promoted The occurrence and spread of computer vulnerabilities and viruses have caused endless network security problems.Facing the continuous emergence of network security problems,although traditional intrusion detection methods have certain attack defense capabilities,they can no longer meet the current requirements for network security defense.At present,there are many methods for detecting malware,but because the types of malware attacks will continue to upgrade and change,and the traditional anomaly detection methods cannot detect unknown attack types,it is difficult to detect the latest anomalous attack behavior.Anomaly detection technology based on artificial intelligence methods such as neural networks,due to its own learning adaptability and dynamic monitoring capabilities,has qualitatively improved network security technology,not only can protect user data,but also can use artificial intelligence methods for automation The program detects abnormal attack behaviors and responds in a timely manner.In addition,it can use its model for early warning detection and discover potential abnormal behaviors.Because of this,this anomaly detection technology based on neural networks has become the focus of attention in the field of network security.This paper first summarizes the current situation of network security problems in the cloud service environment,analyzes the existing network security problems,and then leads to the related intrusion detection technologies.Then it combines the anomaly detection technology which is used to detect the deviation between acceptable behavior with cloud service characteristics,and designs and builds an anomaly detection system model based on BP neural network algorithm Type.The main research work of this paper is as follows:(1)The related technologies about anomaly detection are studied in detail,and then by comparing the advantages and disadvantages of these technologies,an anomaly detection system model based on BP neural network algorithm adopted in this subject is proposed.(2)Focus on the research of BP neural network algorithm,which studies the BP algorithm principle and specific derivation steps in detail,and then analyzes the problems of the traditional BP algorithm and proposes corresponding improvement measures to optimize its defects to a certain extent improve.(3)Build an anomaly detection system model based on the BP neural network algorithm,and describe the overall architectural arrangement of the model in detail,and then extract the event information from the unstructured log information after preprocessing such as data cleaning and mode conversion,as features,Then it is applied to the built model for training and testing.Finally,the evaluation performance of the model is evaluated in terms of detection accuracy,missed report rate,time overhead and other evaluation indicators.The test results show that the model is correct in detection and missed report rate in terms of performance.