Final assembly quality prediction method of P Company based on machine learning
基于机器学习的P公司总装质量预测方法

摘要

机器学习作为一个近期很热门的话题，被越来越多的研究所应用。为了给P公司总装质量性管理提供更好的技术支持，提出了一种基于深度置信网络（DBN）的产品质量预测模型，可通过产品装配历史数据的支持，结合现场的变化情况，判断出产品的合格率情况。以P公司某产品2019年的生产数据作为初始样本，经影响因子提取、去除噪声和筛选稳定点等处理后，形成最终样本集。利用 DBN 网络对样本数据进行特征提取，建立P公司某产品质量预测模型。结果表明：通过测试集验证，基于参数调优的 DBN模型相能准确地预测状态参数，拥有优良的稳定性。

Abstract

As a hot topic recently, machine learning has been applied by more and more researchers. In order to provide better technical support for the quality management of the final assembly of P Company, a product quality prediction model based on deep confidence network (DBN) is proposed. The product qualification rate can be judged by the support of the product assembly history data and the change of the site. The 2019 production data of a product of Company P is taken as the initial sample, and the final sample set is formed after processing such as impact factor extraction, noise removal and stable point screening. The DBN network is used to extract the features of sample data, and a product quality prediction model of P Company is established. The results show that the DBN model based on parameter tuning can accurately predict the state parameters and has good stability through test set verification.