计算机与数字工程2019，Vol.47Issue(5)：1199-1202,1269,5.DOI:10.3969/j.issn.1672-9722.2019.05.036
一种具有自适应动量因子的BP神经网络算法实现
A Novel Collaborative-Filtering Recommendation Algorithm Based on Adaptive Momentum Factor BP Algorithm
摘要
Abstract
The rapid development of Internet technology have given people a lot of data,massive data information greatly af?fectes the consumer's election and purchase efficiency. Traditional collaborative filtering based on BP neural network will lead to lo?cal mininum and slow convergence. In this paper,by improving the BP neural network algorithm,the slow convergence of the prob?lem is optimized. The solution is to increase the momentum on the basis of the momentum of the variable to be adaptive,so that it can change in real time,the result of instability caused by weight adjustment shock is solved well. It is proved that the improved al?gorithm greatly improves the network training efficiency and improves the recommendation efficiency.关键词
改进的BP神经网络/自适应动量因子/协同过滤Key words
improved BP neural network/adaptive momentum factor/collaborative filtering分类
信息技术与安全科学引用本文复制引用
杨文龙,肖程望..一种具有自适应动量因子的BP神经网络算法实现[J].计算机与数字工程,2019,47(5):1199-1202,1269,5.
