Special line traffic early warning method and device based on traffic early warning model, and storage medium

Abstract

The invention discloses a special line traffic early warning method and device based on a traffic early warning model and a storage medium, and the method comprises the steps: collecting a data source, and carrying out the data integration processing, and obtaining a user-based traffic wide table with complaint feedback information and fault declaration records; eliminating the hours of 'zero traffic' in the traffic wide table to obtain an effective internet surfing hour list of the user; determining a traffic full early-warning threshold value, and screening out users who feel network access painful; outputting an early warning user list in a graded manner according to the proportion of the traffic full time and the effective internet surfing time; carrying out cross validation according to the traffic statistical condition, the user data, the complaint feedback information and the fault declaration record in the early warning user list and the traffic wide table, and carrying out early warning user service and customer maintenance in a targeted manner. According to the invention, the analysis blank of the flow use condition during the current special line customer service and maintenance can be filled up, and accurate data support is provided for the customer special line service, customer marketing and contract extension.

Description

technical field [0001] The invention relates to the technical field of data modeling based on network traffic characteristics, in particular to a dedicated line traffic early warning method, device and storage medium based on a traffic early warning model. Background technique [0002] With the development of Internet private line services, major operators need to carry out refined analysis and early warning of private line traffic. On the one hand, operators find that there is a trend of "full traffic", which may cause failures due to insufficient speed such as slow Internet access and Internet freezes. Therefore, marketing measures must be taken in advance to reduce leased line failures and improve customer perception and online experience; Customers are maintained accordingly to reduce the rate of customer churn. Therefore, it is an urgent problem to be solved to construct an Internet private line traffic early warning model for private line traffic early warning analysi...

AI Technical Summary

Problems solved by technology

However, judging the peak flow rate within a period may lead to inaccurate judgment. On the one hand, due to the collection of the 5-minute cycle flow of the port by the equipment network management, there are peak glitches, that is, there are very few collected flow values, which are different from the real values. difference, and taking the peak flow rate in the period, it is likely to get these burr flows, thereby amplifying the influence of these distorted burr flows; There is no scientific basis for judging whether the peak flow rate reaches the opening rate, which affects the customer's online perception. Regarding the user's private line opening rate, whether the actual network traffic is not used enough, the lack of a scientific model as a basis for judgment, resulting in the defect of low prediction accuracy

Images