基于改进的遗传算法优化BP神经网络并用于红酒质量等级分类

Improved GA-BP Network Used to Wine Quality Classification 

摘要:

针对以往的红酒生产过程中,红酒质量分类过程复杂且低效,因此研究一种高效可靠的智能分类识别方法很有必要;这里在红酒的多种物理化学成份测定的基础上,使用人工智能理论中的神经网络构建分类模型,实现对红酒质量的高效分类;并用改进的遗传算法对BP神经网络中的缺陷做了一定的优化;对比传统BP网络分类效果,结果表明,改进后的神经网络收敛速度更快,各个质量等级的分类正确率均提高了10%左右;对红酒加工企业具有积极的实际参考价值。

Abstract:

In the past, the quality of red wine classification process is complex and inefficient.So it’s necessary to study a highly efficient and reliable classification method.Here,on the basis of determination of various physical and chemical composition, using the theory of artificial intelligence to construct classification model.Implementation of the efficiency of the wine quality classifying.We also improve the conventional neural network.The results show that the improved neural network work with high efficiency.It have positive practical reference value to processing enterprises.