Design of Massive Data Management System Based on DBN Algorithm

With the rapid development of the network, the World Wide Web has become the carrier of a large amount of information, which contains a lot of valuable information, and its great potential can not be underestimated. At present, both at home and abroad are paying attention to the potential business opportunities brought by big data, and all walks of life have their own potential business data to achieve the greatest commercial profits in the market. At present, data processing technology in storage system has its own advantages, but it also has its own shortcomings in large-scale data processing. This article takes a large-scale transaction processing system as the research platform, and designs and implements a massive database monitoring and management system to address the main problems in massive database monitoring and management. The system architecture, running model, module design, and key technologies involved are included. Then, using labeled data, the algorithm based on L-BFGS is used to fine tune the system model. At the same time, in order to avoid overfitting, the drop out method is introduced in the pre training and fine tuning process. The experimental results using five fold cross validation and different scale datasets show that the proposed classification method outperforms KNN, SVM, and BP algorithms, and the classification results are relatively stable.