论文部分内容阅读
水中氟离子浓度超标带来的水污染是令人关注的热点问题。本工作运用遗传算法-支持向量回归筛选影响钴铝层状双金属氢氧化物(Co-Al LDHs)氟离子吸附容量的主要特征变量,结果表明复合材料的钴元素的摩尔比(Co%)、吸附剂的剂量(M)、溶液的pH值(pH)、溶液的氟离子浓度(C)是主要特征变量。利用上述特征变量构建支持向量回归模型,留一法交叉验证的均方根误差和平均相对误差分别为0.501和19.5%,实验值和预报值的相关系数为0.943。设计样本基于支持向量回归模型(SVR)预报的氟离子吸附容量与验证实验结果相一致。因此,本工作所建立的支持向量回归模型有望在氟离子吸附容量预报工作中得到进一步的应用。
Water pollution caused by excessive fluoride ion concentration is a hot issue of concern. In this work, genetic algorithm-support vector regression was used to screen the main characteristic variables affecting the fluoride ion adsorption capacities of cobalt-aluminum layered double hydroxide (Co-Al LDHs). The results showed that the molar ratio of cobalt (Co% The adsorbent dose (M), solution pH (pH), solution fluoride ion concentration (C) are the main characteristic variables. The support vector regression model was constructed by using the above feature variables. The root mean square error and average relative error of the left one-way cross-validation were 0.501 and 19.5% respectively, and the correlation coefficient between the experimental and forecast values was 0.943. The fluoride ion adsorption capacity of design samples based on SVR prediction is in agreement with the validation experiment. Therefore, the support vector regression model established in this work is expected to be further applied in the prediction of fluoride ion adsorption capacity.