首页> 外文期刊>Phosphorus, Sulfur, and Silicon and the Related Elements >Application of neural networks with Bayesian regularization for modeling in the synthesis of hydroxyapatite processed from calcium carbonate and phosphoric acid [Application des réseaux de neurones avec la regularisation bayesiénne pour la modélisation de la synthèse de l'hydroxyapatite élaborée à partir du carbonate de calcium et de l'acide phosphorique]
【24h】

Application of neural networks with Bayesian regularization for modeling in the synthesis of hydroxyapatite processed from calcium carbonate and phosphoric acid [Application des réseaux de neurones avec la regularisation bayesiénne pour la modélisation de la synthèse de l'hydroxyapatite élaborée à partir du carbonate de calcium et de l'acide phosphorique]

机译:贝叶斯正则化神经网络在建模中由碳酸钙和磷酸合成羟磷灰石的应用磷酸]

获取原文
获取原文并翻译 | 示例
           

摘要

We have used a new, robust model mapping techniquea Bayesian-regularized neural networkto develop a quantitative relationships model for the synthesis of the phosphocalcic hydroxyapatite by precipitation from a calcium carbonate solution and a phosphoric acid solution. This model was preformed by using a set of factors consisting on the pH of reactional medium, the Ca/P molar ratio of the reagents, reaction time, and the initial concentration of calcium. The results show that the method is robust and gives satisfied results. The Levenberg-Marquardt's algorithm implemented in the neural network Matlab's toolbox allowed a drastic improvement of the performance of the model. Very satisfactory results are then obtained by testing the validity by cross-validation technique. We have also turned our interests to the explanatory capacities of our methodology to explore the relative contribution and/or the contribution profile of the input factors by using Garson weight portioning method.
机译:我们已经使用了一种新的,健壮的模型映射技术(贝叶斯正则化神经网络)来开发一种定量关系模型,用于通过碳酸钙溶液和磷酸溶液中的沉淀来合成磷酸钙羟基磷灰石。该模型是通过使用一组因素来建立的,这些因素包括反应介质的pH值,试剂的Ca / P摩尔比,反应时间和钙的初始浓度。结果表明,该方法是鲁棒的,并给出满意的结果。在神经网络Matlab的工具箱中实现的Levenberg-Marquardt算法可以极大地改善模型的性能。然后通过交叉验证技术测试有效性获得了非常令人满意的结果。我们还把兴趣转向了我们的方法的解释能力,即通过使用Garson权重分配法来探索输入因子的相对贡献和/或贡献曲线。

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