Research on Anomaly Data Recognition Method Based on Deep Confidence Network

This study aims to address the difficulty of identifying abnormal data samples in cloud platforms by constructing a new anomaly data recognition method based on deep confidence network ensemble learning algorithm. The new method utilizes ensemble learning algorithm to solve the problem of imbalanced samples during data operation, and improves the traditional ensemble algorithm using temporal key indicators, Finally, a new adaptive threshold voting algorithm is proposed to improve the classifier of ensemble learning algorithms and improve the generalization of algorithm performance. The experimental results show that the improved algorithm has the highest sample accuracy of 99.52% and a recall rate of 99.29%. The accuracy of the improved algorithm is 0.85% higher than that of the support vector machine algorithm and 1.79
^%
higher than that of the clustering algorithm. The improved algorithm has a higher accuracy than traditional algorithms and performs better in handling abnormal data.