一种改进的BP算法在消费水平中的应用

通过对标准BP算法的改进,提出了一种L-M贝叶斯正则化优化算法,并把它应用到成都市居民消费水平预测中.经试验验证,L-M贝叶斯正则化的BP神经网络比相同条件下另外两种改进算法有更强的泛化能力,对居民消费水平有很好的预测效果.

Application of a Kind of Improved BP Algorithm to Consumption Level

Through improving standard BP algorithm,this paper proposes a kind of L-M Bayesian regularization optimization algorithm and applies this algorithm to the prediction of Chengdu resident consumption level.Experiment shows that BP neural network of L-M Bayesian regulaarization hais stronger generalization than another two kinds of improved algorithms under the same condition and has better forecasting effect on resident consumption level.