论文部分内容阅读
依据近红外光谱(NIR)产生原理,提出了粒子群优化极限学习机(PSO-ELM)算法,运用于小样本氨水浓度定量分析。通过优化极限学习机(ELM)隐藏节点参数,解决了极限学习机由于输入权值和隐含层偏差随机产生的建模结果具有随机性的问题,提高了预测模型的稳定性、精确度和泛化性能。经实验验证,优化后的PSO-ELM相比ELM,模型预测集均方根误差由0.01166减小至0.00322,预测集相关系数由0.9951提高至0.9979。将优化后的模型预测结果与支持向量机(SVM)、BP神经网络算法等传统方法的建模结果进行对比,优化后的PSO-ELM算法具有较高的精确度和良好的泛化性能,模型预测效果优于传统的定量回归分析算法。
Based on the theory of near infrared spectroscopy (NIR) generation, particle swarm optimization limit learning machine (PSO-ELM) algorithm is proposed and applied to the quantitative analysis of ammonia concentration in small samples. By optimizing ELM hidden node parameters, the stochastic problem of modeling results of the limit learning machine due to the input weight and the hidden layer bias is solved, which improves the stability, accuracy and generality of the predictive model. Performance. Compared with the ELM, the RMSE of the model predictive set is reduced from 0.01166 to 0.00322 and the correlation coefficient of the prediction set increases from 0.9951 to 0.9979. The optimized model prediction results are compared with those of traditional methods such as support vector machine (SVM) and BP neural network algorithm. The optimized PSO-ELM algorithm has high accuracy and good generalization performance. The model The prediction effect is better than the traditional quantitative regression analysis algorithm.