论文部分内容阅读
基于近红外光谱技术对烟煤水分分析的快速、无损性。采集了100个烟煤样品,分成验证集和预测集,验证集85个,预测集15个。利用主成分分析对烟煤的近红外光谱数据进行压缩,然后以主成分为输入,采用偏最小二乘回归建立烟煤水分预测模型。烟煤水分平均绝对相对误差为0.0728,表明该方法用于预测烟煤水分含量是可行的。
Rapid and Nondestructive Analysis of Bituminous Coal Moisture Based on Near Infrared Spectroscopy. 100 samples of bituminous coal were collected and divided into verification sets and prediction sets. There were 85 verification sets and 15 prediction sets. The principal component analysis (PCA) was used to compress near infrared spectra of bituminous coal, then the main component was used as input, and the prediction model of bituminous coal moisture was established by partial least squares regression. The average absolute relative error of bituminous coal moisture is 0.0728, indicating that this method is feasible for predicting moisture content of bituminous coal.