Risk Assessment of Substation based on Hermite orthogonal basis Feedforward Neural Network data Fusion algorithm

A data fusion algorithm of Hermite orthogonal basis feedforward neural network based on MapReduce framework is proposed. The experiment takes the substation experimental data, patrol data and monitoring data as the basic data, and carries on the substation risk assessment based on the data fusion to detect the effect of data fusion, and carries on this with the BP neural network algorithm. The experimental results show that the fusion speed of the proposed algorithm is 2-3 times faster than that of BP neural network algorithm, and the standard error and average absolute percentage error are smaller, that is, the accuracy of substation risk assessment is more accurate, which solves the shortcomings of slow convergence speed, local minimum and network uncertainty of BP neural network.