An improved BP neural network algorithm for competitive sports evaluation

Yongshan Peng

Abstract

BP neural network algorithm is easy to get a local optimum and fall into local loop in calculation, which makes slow training speed and unstable calculation, so the development and application of the algorithm are restricted. This paper uses genetic algorithm to improve the generalization ability of ordinary BP algorithm to overcome the above problems. After analyzing the root causes of the defects of BP algorithm, specific calculation steps of genetic algorithm is improved when used to improve BP algorithm first. Then the calculation flows of new algorithm are redesigned. Finally the improved algorithm is used in region competitive sports evaluation and the experimental results show the superiorities of the improved algorithm. The superiorities include simple algorithm process, fast convergence speed, get out local minimum easily, small oscillation and so on.