Research on Fault Prediction Method for Cloud Environment Based on Information Intelligence Fusion
Zhihong Li, Zhipeng Feng, Jiangang Sun, Danni Shao, Dongxia Bai, Jinyi Sun · 2024
With the rapid development of cloud computing technology, the scarcity of single information acquisition in the event of cloud environment failure leads to the failure of fault discrimination. o solve the above problems, an intelligent fusion technique of information acquisition is proposed. The online operation mechanism of this technique is based on a priori knowledge and deep Boltzmann machine (DBM) model. After classifying and processing multi-source information, dynamic data and static data are acquired. The feature sets of different categories of data are extracted by data-driven extraction, and the Kalman filtering algorithm is used to remove the redundant features in the feature sets and complete the consistent feature description of the information, so as to obtain the information fusion results. The test results show that: the technology has good application performance; the test results of Davidson Boulding Index (DBI) are below 0.017, which can effectively deal with the abnormal data in dynamic data and static data; the results of coefficient of variation are below 0.02. The information fusion results obtained by using this technique can reliably predict the information in the fault state, and the application of this technique is effective.