• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>International Journal of Electrochemical Science >Prediction of Zeta Potential of Decomposed Peat via Machine Learning: Comparative Study of Support Vector Machine and Artificial Neural Networks
【24h】

Prediction of Zeta Potential of Decomposed Peat via Machine Learning: Comparative Study of Support Vector Machine and Artificial Neural Networks

机译：基于机器学习的泥炭分解Zeta电位预测：支持向量机与人工神经网络的比较研究

获取原文
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Zeta potential is crucial for practical applications in electrochemistry. However, the precisedeterimination of zeta potential of decomposed peat is complex and has high requirements to relatedinstructments. Previous study shows that zeta potential of decomposed peat can be predicted by back- propagation (BP) neural network. However, it lacks available comparisons and neglects the importanceof the decomposed stages of peat and the required training times. Here, to extend this research, wepropose a series of novel machine learning techniques including support vector machine (SVM) andartificial neural networks (ANNs) to predict the zeta potential of decomposed peat. Four indicatorsincluding hydrated radius (nm), cation valence, concentration (mol/L) and pH are set as independentvariables while zeta potential (mV) is set as the dependent variable. The SVM, general regressionneural network (GRNN) and multilayer feed-forward neural networks (MLFNs) are developed indifferent decomposed stages, including the slightly decomposed peat, the highly decomposed peat andall decomposed peat. Results show that separating the models based on the decomposed stages havebetter prediction results than taking all decomposed peat in one model. During our studies, the SVM isthe best model for the prediction to the slightly decomposed peat (RMS error: 2.37, training time: 1s),while the GRNN is the best model for the prediction to the highly decomposed peat (RMS error: 2.20,training time: 1s).
机译：Zeta电势对于电化学中的实际应用至关重要。但是，对分解泥炭的ζ电位的精确测定是复杂的，并且对相关指导要求很高。先前的研究表明，可以通过反向传播（BP）神经网络预测分解泥炭的Zeta电位。然而，它缺乏可用的比较，并且忽略了泥炭分解阶段和所需训练时间的重要性。在这里，为了扩展这项研究，我们提出了一系列新颖的机器学习技术，包括支持向量机（SVM）和人工神经网络（ANN），以预测分解泥炭的Zeta潜力。包括水合半径（nm），阳离子化合价，浓度（mol / L）和pH在内的四个指标设置为自变量，而ζ电位（mV）设置为因变量。支持向量机，通用回归神经网络（GRNN）和多层前馈神经网络（MLFN）在不同的分解阶段得以发展，包括略有分解的泥炭，高度分解的泥炭和所有分解的泥炭。结果表明，与将所有分解泥炭纳入一个模型相比，基于分解阶段分离模型具有更好的预测结果。在我们的研究中，SVM是预测稍有分解的泥炭的最佳模型（RMS误差：2.37，训练时间：1s），而GRNN是预测高度分解的泥炭的最佳模型（RMS误差：2.20，训练时间：1s）。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号