基于标准数据集的粘土类矿物样本辨识模式.doc

基于标准数据集的粘土类矿物样本辨识模式 【中文摘要】对于粘土类矿物的辨识,利用X射线粉末衍射技术,进行物相定性分析是一种有效的方法。将收集到的标准X射线粉末衍射粘土类矿物数据,组成一个数据集,再将待测样本的d-I衍射数据,看作是酉空间中的复向量,与标准数据进行比较,从而将物相辨识转化为复向量间的相似性匹配题目,终极得到样本的回属结果。在混合物物相分析方法中,Fiala建立的因子模型方法,推理严谨、方便易行。但在应用过程中存在一些困难其中特征值判据,存在估计值偏大的题目,依据化学因子分析中的误差理论,引进标准差估计判据、约化特征值判据替换原判据,效果较佳;在应用过程中,需求解多重多元线性回回方程,相关文献中未清楚给出具体解决办法,造成相应困难,将该求解题目转化为二次规划题目,再利用微软Office Excel中规划求解模块,从而方便地得到了方程的解,为Fiala的因子模型方法应用,提供了一种途径。基于标准衍射数据集,对样本衍射数据进行分析,从而得到样本回属的辨识模式,是一种可行的研究方法。; 【Abstract】 In all kinds of the clay minerals identified s, the phase qualitative analysis used the X-ray powder diffraction technology is one effective . Collecting the standard X-ray powder diffraction data of mineral of clay, composes a data set; the testing sample\s d-I diffraction data, regards as a complex vector in a unitary space, then comparing with the standard data, thus transs the phase identification as the question of similarity match between two complex vectors, obtains the result of sample’s ownership finally. In mixture phase analysis s, a factor pattern established by Fiala is inference rigorous and convenient applying. But in the applying process, there are some difficulties One eigenvalue criterion, it is relatively large in applying process, based on the error theory of the chemical factor analysis, using standard deviation estimate criterion and the criterion of contract eigenvalue replace the old criterion, the result of verdict is better; In the applying process, it must solve the multiple multi-dimensional linear regression equation, in the related literatures, it has not clearly given the of solution, causes the corresponding difficulty, solving the problem is transed into a quadratic programming problems, then using the Programming Module of Microsoft Office Excel, thus obtained the equation solution conveniently, it has provided one way for the applying of Fiala’s factor pattern . Based on the standard diffraction data set, analyzing the sample diffraction data, thus obtains the sample’s ownership, the identification pattern is one feasible research technique.