Software service quality monitoring method and system based on weighted naive Bayes classifier

Abstract

The invention discloses a software service quality monitoring method and monitoring system based on a weighted naive Bayesian classifier. The method judges whether the QoS of the software belongs to the category of meeting the standard, not meeting the standard or Unable to judge, set the impact factor combination during training, the impact factor is not the impact of the software itself on QoS, calculate the weight of the impact factor combination and prior knowledge, obtain the monitoring results by the classifier during monitoring, analyze, store and return the evaluation to the data Server. The system includes: a controller that collects different quality of service statements, issues data standard instructions required for different quality of service to the observer, transmits the QoS standard that needs to match the data set to the trainer, controls the collection cycle and frequency; database summary analyzer The monitoring results of the software; the service capability evaluation module returns the monitoring results and evaluation results of the software to the data service device; it also includes observers, trainers, optimization sample sets and analyzers.

Description

technical field [0001] The invention relates to a software service quality monitoring method and system based on a weighted naive Bayesian classifier, in particular to the dynamic weighted monitoring of QoS demand indicators during operation, and belongs to the field of information technology. Background technique [0002] Web service technology adapts to any type of Web environment, including Internet, Intranet and Extranet, and realizes the communication between enterprises and enterprises, enterprises and consumers. From the perspective of the key technologies of Web services, Web services still have a lot of research space and challenges. One of the most important issues in user needs is the quality of service (Quality of Service). key to success. Today, when Service-Oriented Architecture is widely used, software systems can dynamically combine some loosely coupled components (that is, Services) with a unified interface definition. However, in complex and changeable Web...

AI Technical Summary

Problems solved by technology

The current probabilistic monitoring methods use hypothesis testing for statistics, and there are also monitoring methods based on Bayesian factors. The former only uses a single probability value for evaluation, and the actual implementability is poor. The error rate of the monitoring results of the latter needs to be reduced. More importantly, Existing methods ignore the influence of the environment on the monitoring results. Different environments will affect our probability monitoring results. These environments include the user's location, network, server CPU, RAM, I / O, etc. For example, users are in different regions. May experience different QoS when using the same service

Therefore, the existing probabilistic monitoring technology has been unable to meet the needs of probabilistic monitoring

Images