It Incident And Change Management Decision Support Method Based On Machine Learning

Incident management and change management are two important processes in the field of IT service management. With the aid of these two processes, IT service system can quickly deal with problems, rapid response to external demand and provide better quality of service. However, with growing customer demands, incident management and change management are suffering with increasing pressure. Facing the huge amount of incident tickets and RFCs, the operator is difficult to guarantee the timeliness and determine the inherent risks. This will lead to a drop in the quality of service or a service interruption. As a result, it is of great value to rapid handle incident and change with automated decision support method.With analysis and study of current research achievements, this paper proposed a decision support method which based on machine learning. And this method also considered with current technical defects and the characteristics of decision support field. The main research contents are as follows:1) Decision support method which based on machine learning: Existing decision support methods have poor robustness of data and high computational complexity, which lead to more resources and time requirement in preprocessing and training. To solve this problem, this paper uses incident ticket and request for change as the basis of decision support, imports four types of features which considering with characteristics of processes. By introducing relevant configuration items and deployment architecture graph, this method can better illustrate incident and change which could improve the accuracy of decision support. Then, this paper provides comprehensive analysis on the learning styles and the specific training algorithms of machine learning. Finally, this paper provides a decision support method which satisfies the requirement of high accuracy, high robustness, and low training time.2) Automated decision support method:Due to the existing process automation solutions still needs a lot of human participation in decision support field, this paper also provides automated decision support method which according to the characteristics of decision support. By introducing the correlation score, deployment architecture graph similarity score, features signature and etc., this paper provides automation process which suitable for decision support. With this method, IT service system could avoid high personnel training cost and maintenance cost while increasing accuracy and timeliness in processing incidents and changes.Finally, this paper gives a simulation in the change management. And the result illustrates that the automation decision support method, this paper proposed, does better than the existing decision support methods in decision accuracy and the training time.