摘要
介绍了应用在毛精纺面料织造质量预报过程中的人工神经网络技术(ANN)和多元线性回归方法,给出了2种方法在建立各自模型时的主要工作,并在此基础上建立了织机效率和织疵公分数的2种预报模型,最后通过2种模型的预报结果对比验证了ANN模型和多元线性回归预报模型在毛精纺织造过程预报中的性能,同时得出了2种预报模型在解决线性和非线性问题上的优劣,以及毛精纺织造过程中的纱线品质和工艺参数与织造质量指标之间的线性或非线性关系。
The neural network technology and multiple linear regression used to weaving process of wool worsted were introduced in this paper. The main jobs for setting up two kinds of forecast models by these two methods for weaving process were put forward. The models to predict the loom efficiency and fabric defects were given, too. The character of the neural network and multiple linear regression models were demonstrated and contrasted by the forecast results, and the contrast between these two models for solving the nonlinear and linear problems was shown. The linear or nonlinear relationships between the yarn quality and weaving parameters and woven quality indexes were also given.
出处
《毛纺科技》
CAS
北大核心
2008年第7期42-44,共3页
Wool Textile Journal
关键词
毛精纺
织造
质量预报
人工神经网络
多元线性回归
wool worsted
weaving
quality forecasting
artificial neural network
multiple linear regression
