数控加工铣削试验数据的清洗与处理

样本数据作为神经网络模型训练的导师,其质量直接影响了神经网络的学习能力以及神经网络模型的预测能力。以DMC60H为试验平台,以壳体类铝合金零件加工为研究对象,提取数控铣削加工试验数据;通过对数控铣削参数试验数据的分析与研究,提出了试验数据与样本数据的处理原则,分析了验证数据的构成以及所占数量。实现了样本数据的优化,并同时剔除了样本数据中的错误信息。

The cleaning and processing of CNC machining milling test data

As a neural network model trained instructors are becoming more widely appreciated,sample data directly affects the quality of the learning ability of neural networks and neural network model predictive power.Take the Tokushima Magee 60H as the test platform,Shell aluminum alloy parts processing as object to extract experimental data for NCN milling.Processing principles of the test and sample data are proposed through the analysis and studying of the CNC milling parameter optimization.Analysis to verify the number of data composition and proportion.So we achieved the optimization of sample data and removed the sample data in the error message.