Large-scale Data Mining Algorithms and Models for Equipment and Power Supply Service Information Fusion

With the rapid development of computer technology, the requirements for power quality of power systems are becoming more and more stringent. This article aims to study key issues in the integration of information resources for power supply services. First, this article provides an analysis of current device and communications data sources and their characteristics. Secondly, this article introduces feature selection data mining methods such as multi-user multipath ensemble algorithm, model-based feature extraction strategy and heuristic matching strategy based on task environment. Finally, this article uses Matlab for verification and obtains experimental results. The accuracy of this algorithm is at least 90% and can reach as high as 99%; the memory usage of the model is within 0.06; the stability of the model is between 0.8-0.89.