Flow prediction model training method and device, flow prediction method and device, equipment and medium

A traffic forecasting and training method technology, applied in the network field, can solve the problems of traffic peak hysteresis accuracy and low accuracy, achieve the effect of solving hysteresis and low accuracy, ensuring diversity, and improving prediction accuracy

Active Publication Date: 2020-12-22
MAIPU COMM TECH CO LTD
View PDF13 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The model trained through the above implementation process can combine the processing results of the sample data characteristics of the previous traffic with the gated recurrent neural network and the sample attribute characteristics during prediction when performing traffic prediction, so that it can be used in traditional gating On the basis of cyclic neural network prediction, combined with more abundant flow sample attribute characteristics, the prediction of flow is no longer based solely on the flow evolution law at historical moments, but will combine different attribute characteristics of different flows Make predictions so that the prediction results are associated with attributes at different moments to be predicted, so as to solve the problems of hysteresis and low accuracy in the prediction of traffic peaks

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Flow prediction model training method and device, flow prediction method and device, equipment and medium
  • Flow prediction model training method and device, flow prediction method and device, equipment and medium
  • Flow prediction model training method and device, flow prediction method and device, equipment and medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] At present, in order to realize the prediction of traffic, gated cyclic neural network (such as LSTM, GRU, etc.) is often used as the prediction model, and the historical network traffic before the prediction time is processed by using the trained gated cyclic neural network. to get the predicted value.

[0044] However, in practical application scenarios, there are often many unexpected situations in the use of network traffic, and the stability of network traffic data changes is not strong. The problem of lag and low accuracy. For this reason, a new traffic forecasting model is provided in the embodiment of this application, which can be found in figure 1 As shown, it includes: gated recurrent neural network and fully connected network.

[0045] Among them, the input of the gated recurrent neural network is the sequence formed by the network traffic. The sequence is processed through a gated recurrent neural network, which outputs a memory state.

[0046] In the e...

Embodiment 2

[0121] On the basis of the first embodiment, this embodiment takes the LSTM network and the GRU network as examples to further illustrate the solution of the present application.

example 1

[0122] Example 1, the gated recurrent neural network is an LSTM network, including the following steps:

[0123] Step 301), aggregating the sampled original network traffic into network traffic at preset intervals to obtain a training set.

[0124] Step 302), using the "isolation forest" algorithm to detect abnormal network traffic in the training set.

[0125] Use the algorithmic formula S(x,n) = 2 –(E(h(x))) / c(n) The abnormal score corresponding to each network traffic in the training set is obtained, and it is judged whether the abnormal score is greater than the preset score threshold, so that the network traffic with the abnormal score greater than the preset score threshold is determined as abnormal network traffic.

[0126] In the embodiment of the present application, the preset score threshold may be 0.5.

[0127] After the abnormal network traffic is detected, the average value among neighbors is used to replace the detected abnormal network traffic. For example, ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a flow prediction model training method and device, a flow prediction method and device, equipment and a medium. The flow prediction model training method comprises the steps: obtaining a flow data set, intercepting the flow data set through a sliding window operation, and obtaining the sample data features of the flow of the size of a sliding window and the sample labels corresponding to the sample data features of the flow; obtaining a next flow value after the sample data characteristics of the sample label flow; obtaining sample attribute characteristics corresponding to the sample label; processing the sample data characteristics of the flow by using a gated recurrent neural network of the model, and inputting the output memory state and the sample attribute characteristics into a full-connection network of the model for regression prediction; and according to the sample label and the prediction result, updating the parameter weight through a back propagation algorithm to obtain a trained flow prediction model. Therefore, a prediction result is associated with the attributes of different flows, and the problems of hysteresis quality and low accuracy of flow peak prediction are solved.

Description

technical field [0001] The present application relates to the field of network technology, in particular, to a traffic prediction model, training method, prediction method, device, equipment and medium. Background technique [0002] With the popularization of the network, the scale of network traffic is constantly updated, and efficient and reasonable allocation of network resources has become particularly important. On the one hand, the unreasonable allocation of network resources may cause some network resources to be unable to be used normally due to exhaustion, and even cause network paralysis; while another part of network resources is in a state of excess, which seriously affects the user's online experience. On the other hand, although the network allocates network resources reasonably in the early stage, the network traffic is sudden. At this time, the originally sufficient network resources may be in short supply. To solve this problem, the existing SDN (Software D...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/24G06Q10/04G06N3/08G06N3/04G06K9/62
CPCH04L41/147H04L41/145G06N3/084G06Q10/04G06N3/045G06F18/214
Inventor 徐海兵郭久明
Owner MAIPU COMM TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap