A Predictive Method for Enterprise Market Competitiveness using LR-GWO-BP Neural Network

This paper introduces an innovative approach to enhance the predictive accuracy of BP Neural Network. Leveraging the Grey Wolf Optimization algorithm, the proposed methodology refines the initial weights and thresholds of the BP Neural Network, addressing traditional GWO limitations. To overcome diminishing accuracy issues during iterations, the LR-GWO algorithm is introduced, featuring a segmented adjustable decay factor and a probability-based selection mechanism for leadership-level grey wolves. This strategic enhancement significantly strengthens the algorithm's global and local optimization capabilities. In practical application, the LR-GWO-BP neural network proves effective in predicting market competitiveness within enterprises. Experimental results demonstrate a consistent reduction in overall error, highlighting the potential of this novel approach in transforming predictive modeling within enterprises. The combination of BP Neural Networks with the optimized LRGWO algorithm emerges as a solution for predictive modeling.