Internet Traffic Behavior Analysis And Application Based On Log Minin

With the development of information technology,human life is inseparable from the network.People’s daily activities on the network will produce a huge amount of log data,which reflects the user’s behavior information on the network.How to process these logs and extract some valuable information hidden behind them,so as to analyze the user’s behavior habits,exceptions and other information has become a hot field for many researchers in recent years.Enterprise level online behavior monitoring equipment can record the network behavior of all users.The traditional management mode is to check the log when there is a fault or problem.This way can only passively solve the problem,not prevent the problem.Consider mining and analyzing the log data we usually don’t pay much attention to through the data mining method based on machine learning,and actively understand the daily online behavior habits,preferences and abnormal information of employees and users,so as to provide data basis for the strategy adjustment of our daily network management.Based on the above actual demand scenarios,this paper starts the research in this area.Firstly,the network user behavior theory and the mining technology theories such as Apriori algorithm,BP neural network and sparrow search algorithm are described in detail;Then the data is collected through the database system,and the collected data is preprocessed;Then,the processed network traffic data are statistically analyzed.By establishing the correlation analysis model based on Apriori algorithm,the correlation of employees’ online application traffic is analyzed,and the current network traffic control strategy is studied and analyzed according to the correlation results,so as to put forward strategic adjustment suggestions.The suggestions have been tested in the actual environment and achieved good bandwidth optimization results;Finally,by constructing the BP neural network model based on sparrow search algorithm,the online traffic is modeled and analyzed according to the time period.The research results are applied to predict the network traffic of employees.From the experimental results,the optimized BP neural network model can achieve good prediction accuracy under the condition of low sample size,which shows that the BP neural network model optimized by sparrow search algorithm is an effective analysis and prediction model of users’ online behavior.