Wireless network flow prediction method based on local minimax probability machine

Abstract

The invention provides a wireless network flow prediction method based on a local minimax probability machine, which comprises the following steps of: calculating time delay of a time sequence by using a mutual information algorithm; calculating an embedding dimension by using a Cao method; carrying out phase space reconstruction on the time sequence by using a delay coordinate state phase space reconstruction method to obtain a phase space; calculating the number of proximal points of the phase space by using an AICi information criteria; constructing a proximal point set by using a KD-Tree-based proximal point selection algorithm; rejecting pseudo proximal points from the proximal point set by using a Ki strategy to obtain a cut proximal point set for constructing a training characteristic set of the minimax probability machine; regulating parameters of the minimax probability machine; and training a model to obtain a prediction value. The wireless network flow prediction method disclosed by the invention realizes the localization of the prediction model of the minimax probability machine and can carry out wireless network flow prediction more accurately and more timely.

Description

technical field [0001] The invention relates to the field of wireless networks, in particular to a method for predicting wireless network traffic. Background technique [0002] In recent years, with the continuous promotion of wireless network technology applications, security issues have become one of the most critical issues encountered in the development of wireless local area networks. In addition to access control technology based on key management and authentication, network traffic prediction and abnormal Detection has gradually become an important means to solve the above problems. At the same time, the analysis and prediction of network traffic is of great significance to network planning and network resource management. [0003] Prediction is to analyze and predict the future state of something based on the current development trend and laws. It includes collecting historical data and using a certain mathematical model to solve and extrapolate future data. Networ...

AI Technical Summary

Problems solved by technology

But the Euclidean distance has flaws in measuring similarity: even a slight movement of two very similar vectors on the time axis may cause the Euclidean distance between the two to become very large

Disadvantages: When determining the number of NNPS and how many false neighbors need to be eliminated in the K strategy, this method mainly relies on manual experience, lacks scientific basis and necessary calculations; in addition, the time complexity of the DTW algorithm is also significantly higher than that based on Similarity Calculation Method of Euclidean Distance

However, MPMR, like SVM, has flaws when building a prediction model with real-time computing power: its main idea is to use all historical information to infer the future, so when the input data reaches a certain scale, the amount of calculation will increase sharply, training The time will become very long or even exceed the time interval of data collection, so that the prediction has lost its practical significance

Images