基于人工神经网络的工业供应链销售预测方法

Sales Prediction Method for Industrial Supply Chain Based on Artificial Neural Network

为解决工业供应链中存在的精度低、非智能以及无法处理复杂样本的问题,提出一种基于改进人工神经网络的销售预测方法。以加拿大某机电产品销售公司的真实销售数据作为输入样本,利用基于实验数据改进的人工神经网络进行学习训练,进行销售预测,将结果与未改进的人工神经网络和较先进的卷积神经网络和高斯混合模型以及销售公司的销售数据作比较,从准确率、召回率和F值三个指标分析改进人工神经网络的预测精度。实验结果表明,改进后的人工神经网络在三个指标方面均表现出更好的性能,能够较好地预测销售成交情况。

Aiming at the problems of low precision,unintelligence and inability of processing on complex data in the qualitative sales prediction for industrial supply chain,a method based on improved artificial neural network is proposed in this paper.The actual sales data of a Canadian mechanical and electrical product sales company is used as an input sample,and the improved artifical neural network based on experimental data is used for learning,trainng and sales forecasting.Compared with unimproved artificial neural networks and advanced convolution neural network,Gaussian mixture model and the sales data,the prediction accuracy of the results is analyzed from the three indicators of precision,recall and F value.The experimental results show that the improved artificial neural network shows better performance in all three indicators,and can better predict sales transactions.