Group-oriented Service Reliability Prediction Method in Big Data Scenario

Abstract

The present invention relates to a group-oriented service reliability prediction method under a big data scene. The method mainly includes solving the user similarity in an inherent group, dynamically calculating the similarity threshold, and calculating the service similarity reliability matrix method at low cost. The reliability matrix model is calculated offline, and the calculated reliability matrix is ​​distributed through CDN, and user requests are processed based on geographic location to ensure that user requests can be responded quickly. The invention can not only calculate the similarity between individual user groups, but also use the synergy theory to predict the reliability of the users in the group in the process of using the service.

Description

technical field [0001] The invention relates to a service reliability prediction method, in particular to a group-oriented service reliability prediction method under a big data scene, and belongs to the technical field of prediction systems. Background technique [0002] In recent years, with the wide popularization of Internet technology, some web-based services such as online shopping, online ticket booking, and online real-time news are also popular. However, because the release of network services is more convenient than traditional services and the control is relatively relaxed, unreliable services are increasing in the Internet, and this phenomenon seriously reduces the quality of user experience, which also affects the development of these services. Reliability prediction of these web services has received increasing attention. [0003] At present, there are few researches on the reliability prediction of single service, and most of them study the reliability of com...

AI Technical Summary

Problems solved by technology

[0006] 1) Less consideration is given to service reliability from the perspective of service users, ignoring the impact of user environment and behavior on service reliability, often resulting in lower accuracy of predicted service reliability;

[0007] 2) When predicting service reliability, we ignore the fact that when users with high similarity call the same service, the reliability of the service basically does not fluctuate, thus increasing the consumption of the prediction method

In short, the existing service reliability prediction methods consume a lot of energy and are difficult to apply in real-time big data scenarios

Images