多组分类

Using the methods of classical shape size measuring and chromosome analysis, we studied the morphological characters and Karyotype of Cobitis sinensis in Xiangjiang River. We also investigated the evolution line and Karyotype polymorphism of the C . sinensis by the methods of cytological classification and cytocatalytic. Our results suggest that the formation of chromosome polymorphism was concerned with heterologous hybridization and some of C . sinensis may perpetuate in a clonal manner by natural gynogenesis.

本文以湘江流域的中华花鳅为研究对象，运用经典的形态测量观察方法、染色体制片技术对中华花鳅的形态生态特征、染色体组型进行了初步研究：运用细胞分类和进化方法对中华花鳅的核型演化和核型多态现象的形成进行了分析，提出：其核型多态的形成牵涉到鳅科鱼类的异源杂交，某些中华花鳅可能是行雌核发育生殖的。

In this chapter, three kinds of target recognition methods are performed, which are:①Target recognition method based on the description of polarization parameter plane. The echo polarization states of target are projected onto the polarization state plane described by the ellipticity ε and the tilt angle τ of the polarization ellipse, the change of parameter following ferquency becomes the chart. According to the changing trait of the chart, the multidimensional polarization feature space of target has been contructed. Furthermore, a series of polarization feature parameters used in designing the structure of target recognition device are extracted, and they are insensible to the posture of target.②Target recognition method based on the description of Poincare polarization sphere. The echo polarization states of target expressed by Stokes vector are projected onto the Poincare polarization sphere. The conception of polarization ferquency stability, which is used in describing the dynamic distribution characteristics of the target echo polarization states on Poincare polarization sphere, has been defined. A group of polarization feature parameters used in designing the structure of target recognition device are extracted, and they are insensible to the posture of target.③Target recognition method based on the description of frequency sensitivity. In accordance with the conception of the polarization state distance defined on Poincare polarization sphere, the frequency sensibility of the physical structure property of target has been investigated, the frequency distribution feature curves in PSD domain are obtained, and targets'features are extracted by means of the curve-fitting method with Least Square Criterion.

这章具体研究了基于三种极化散射特性描述的相应的目标识别方法：①基于极化参数平面描述的目标识别方法，将目标回波极化状态投影到以极化椭圆参数，即椭圆率角ε和倾角τ表征的极化状态平面上，参数随观测频率的变化就形成了图，根据图的变化特点构造了目标的多维极化特征空间，并提取了不敏感目标姿态变化的极化特征参数组来设计目标的识别器结构；②基于Poincare极化球面描述的目标识别方法，采用Stokes矢量表征目标回波的极化状态，并将其投影到描述极化状态的Poincare极化球面上，定义了极化频率稳定度的概念用以刻画目标回波极化状态在Poincare极化球面上的动态分布信息，提取了准方位不变性的目标极化特征，最后设计了目标的识别器结构；③基于频率敏感性描述的目标识别方法，通过在Poincare极化球面上所定义的极化状态距离的概念，研究的是复杂目标物理结构特性对探测信号频率的敏感程度问题，获得了在极化状态距离下的频率分布特性曲线，采用最小二乘估计曲线拟合方法，它既用于极化特征的降维，同时又直接将拟合参数作为目标的分类特征。

Moreover, the actual allele frequency of most varieties deviates far from Hardy-Weinberg equilibrium. All PPB、na 、I、h、Gi and Fst have proved to be the references to elucidate that ISSR is a most powerful tool to analyze genetic diversity, compared with the RAPD marker and the allozyme marker is less strong ordinally. We could divided the 70 samples into A, B, C, D and E five groups using three methods according to genetic distance clustering. There is a bit displacement for few varieties in different clustering maps, but the most are similar to morphological analysis despite that there is still a great difference among cultivars in the same one group. The above results imply that the three methods have the different sensitivity and resolution in genetic distance analysis of close varieties. The Mantel test indicates that the results from the three kinds of markers have the significant correlation, which demonstrates that the number of the used three kinds of markers is enough to exactly detect the diversity of all 70 samples to ideal extent. And these methods can be used to evaluate the diversity of the whole group using the miscellaneous samples instead of the individual sample, of the Gerbera jamesonii are mainly from tissue culture plants. In conclusion, the above study results provide a reference for the application of three kinds of molecular markers to molecular marker-assisted breeding of flower. 2. The genetic diversity among the eight introduced cut-flower varieties of Ranunculus asiatica was analyzed by the ISSR markers. Based on the genetic clustering tree, all the colorful flower varieties are clustered into one group, and the white flower varieties into another group. Moreover, among the former group the yellow flower varieties are clustered into one sub-group, and the reddish flower varieties, such as rose color, pink, nacarat, are clusetered into another sub-group.

由三种分子标记的分析结果可以看出，等位基因平均值、观察杂合度、Fis值、Fit值皆较高，表明非洲菊等位基因较丰富，杂合基因偏多，且绝大部份品种的实际等位基因频率在品种内偏离了Hardy-Weinberg equilibrium；PP8、na、Ⅰ、h、Gi及Fst皆表明，ISSR检测遗传多样性的能力最强，其次是RAPD，等位酶最低；根据遗传距离进行聚类，三种方法都把70个品种分成A、B、C、D、E五个大组，每一组中除少数品种发生位移外，大部份品种分类结果相似，且与形态分析结果有相似性，但在每一组中，品种间的聚类差别较大，表明这三种方法在近距离品种间检测遗传变异时灵敏度及分辨力不同；Mantel检测表明，三种标记的分析结果有显著相关性，表明所用的三种分子方法的标记数量已经可以相对无偏地检测到70个品种间遗传变异；非洲菊为组培苗，三种标记的检测结果皆表明，混合样品可以作为个体样品的代表，对整个居群的遗传多样性进行评价；这些研究结果可为三种分子标记方法在花卉分子辅助育种中的进一步应用提供借鉴。

The taxonomy module allows multiple lists of categories for classification and offers the possibility of creating thesauri (controlled vocabularies that indicate the relationship of terms) and taxonomies (controlled vocabularies where relationships are indicated hierarchically).

分类模组允许多分类的列表，同时提供产生词典的可能性（指出次分类关联的词汇表控制），并且支援分类法（控制显示有阶层等级关系的次分类）。

In a database the concept of an example might change along with time which is known as concept drift When the concept drift the classification model built by use of old data is unsuitable for classify new data Therefore concept drift has become a hot issue in data mining in recent years Although many algorithms had been proposed to resolve this problem they can not provide users with the reason of concept drift However a user might be very interested in such rules For example doctors want to find what makes disease change; researchers want to know the reason of the variety of the weather; and decision makers would like to understand why a customer's shopping habit change In this thesis we propose a Concept Drift Rule mining Tree called CDR-Tree to solve this problem CDR-Tree can not only find the rule of concept drift also build the prediction model for both old and new data at the same time

无论在大型资料库或现实生活中，同一资料样本的概念有可能会随著时间的递移而改变，也就是产生所谓的概念漂移。当样本发生概念漂移时，由旧有资料所建构的分类模组将不再适用於预测新获得的资料，因此，近年来概念漂移已成为资料探勘中一项热门的研究议题。虽然已有？多学者提出不同的技术来解决概念漂移的问题，但是这些方法都是利用修正或重建的方式来产生适合新资料的预测模组，并无法提供造成概念漂移的原因。然而对使用者而言，其感兴趣的可能正是这些引起概念漂移的规则，如医生可能想了解引起疾病变化的主因、学者会想要知道气候转变的规则、或是决策者想找出顾客购物习惯改变的因素等。因此，本论文提出概念漂移规则探勘树( Concept Drift Rule mining Tree )，简称CDR-Tree，来解决这个问题。CDR-Tree不但能探测出造成概念漂移的主要原因，亦能同时建立新旧资料的预测模组以供决策者运用使用。

Data mining with decision tree analysis revealed that different combinations of morning stiffness, slight red tongue, joint tenderness and nocturnal polyuria in the CM group, and those of white furin tongue, CRP level, leucocyte count and morning stiffness in the WM group showed different efficacy, whichwere also verified in the randomly chosen verifying set.

决策树分析结果显示：中药治疗组中晨僵、舌淡红、关节压痛程度、夜尿多4项观测指标不同组合患者的中药治疗疗效有差异；西药组中舌苔白、C反应蛋白、白细胞数量和晨僵4项观测指标不同组合患者的中药治疗疗效有差异；同时，决策树分类的结果在随机选取的验证集中也得到了验证。

Based on the split window method, the classification of the ground land cover types, and the given spectral emissivities range in thermal infrared bands for each ground class and the fact that the actual land surface temperatures are the same for all thermal bands of MODIS, the land surface temperature were determined by means of maximum correlation and least square fit.

因此，本文在地表温度遥感定量反演的局地分裂窗算法的基础上，依据MODIS的多个热红外通道，其中的任意两个通道组合，通过分裂窗算法得到的地表温度都应该相等这样一个事实，在对地物进行分类的基础上，给定地物的波段比辐射率的范围，借助于两组温度之间的最大相关和最小方差，通过迭代的方式得到地表温度和分裂窗算法所使用的MODIS热红外通道的波段比辐射率。

Firstly, the features were divided into L groups and the SVM hyperplanes were constructed for each feature of training set. Secondly, the testing set was tested by SVMKNN method, and the decision profile matrixes were obtained. Finally, these decision profile matrixes were implemented by multi-feature fusion method.

在该算法中，设样本集特征可分为L组，先用SVM算法根据训练集中每组特征数据构造分类超平面，共构造L个；其次用SVMKNN方法对测试集进行测试，得到由L组后验概率构成的决策轮廓矩阵；最后将其进行多特征融合，输出最终的分类结果。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。