基于极端学习机的NOx预测模型样本特性研究

针对某燃煤锅炉进行了样本特性实验，选取测试集均方根误差（RMSE）作为性能指标，基于不同测试样本数目分别建立电厂NOx排放的极端学习机预测模型。经过31次实验后的结果表明，随着测试样本数的增加，预测样本RMSE呈增长的趋势。当测试样本数为2时，极端学习机（ELM）可以建立相对准确的预测模型。

Study on the Samples Characteristic of NOx Predicton Model based on Extreme Learning Machine

Under the premise of a certain total sample,the effect of the number of test samples on the prediction performance of NOx is studied.A sample characteristic experiment is conducted for a coal-fired boiler.The mean square error of the test set is selected as the performance index.Based on the number of different test samples,an extreme learning machine prediction model of NOx emissions from the power plant is established.Each experiment is conducted 31 times to analyze sample characteristics.The experimental results show that as the number of test samples increases,the prediction samples RMSE increases.When the number of test samples is 2,ELM can build a relatively accurate prediction model.