结果还指出出现了离群值。
为了创建报告使用的离群值表查询主题。
To create the outlier table query subject used by the report.
离群值与非离群值之间并没有明显的区别。
Thus, there is not a sharp distinction between outlier and non-outlier.
Cognos非常适合分析交互式离群值。
Cognos is very well-suited to support the task of interactive outlier analysis.
这些记录称为离群值。
这个度越高，则该集群中的记录越有可能是离群值。
The higher this degree, the more likely the records in the cluster can be considered outliers.
离群值处理通常不是一项完全自动化的任务。
表CUSTOMER_OL包含关于离群值的相关信息。
The table CUSTOMER_OL contains relevant information about outliers.
您需要一个离群值标记项，以表明一个记录是否为离群值。
You want an outlier-flag query item that indicates whether a record is considered an outlier.
clusterid是记录所属的“离群值”集群的id。
The cluster id is the id of the "outlier" cluster to which the record belongs.
将客户按职业分组，并提供关于每种职业发现多少离群值的概述。
Group customers by profession and provide an overview on how many outliers can be found in each profession category.
因此，一个先进的用户界面和交互模型是成功处理离群值的前提条件。
A sophisticated user interface and interaction model is therefore a prerequisite of successful outlier handling.
并利用其中一种，建立了稳健的多变量离群值的检测方法。
By using one of them, some robust multivariable outlier detection methods are built.
在本节中，学习如何创建允许交互式查看离群值的Cognos报告。
In this section, learn how to create a Cognos report that allows you to interactively inspect outliers.
用Cognosreportstudio创建一个离群值报告。
当样本数据中没有离群值时，这些方法都能得到优良的结果。
When there is no outlier in the sample, these methods can get good result.
概述页面包含一个职业列表，显示每种职业的客户记录数量和离群值数量。
The overview page contains a list of all professions with the number of customer records and the number of outliers per profession.
检验的样本崩溃点是样本中能逆转判决的离群值的最小比例 。
The sample breakdown point of a test for moment estimation of population variance;
离群值详细界面包含被标记为离群值且属于主报告页面选定的职业的客户记录。
The outlier details page contains the customer records that are flagged as outliers, which belong to the profession selected in the main report page.
偏差度高于这个阈值的所有集群被标记为离群值集群，它们的成员都是离群值。
All clusters with an outlier degree above this threshold are marked as outlier clusters and all their members as outliers.
您可以过滤记录，也可以对它们进行排序，从而获得想要查看或必须检查的离群值。
You can either filter the records, or sort them to obtain the outliers you would like to review or that you must check.
首先，如果检查离群值的专家有限，那么可以使用具有最高偏差度的集群的数据记录。
First, if you only have a limited number of experts that are able to check outliers, you simply use the data records that belong to clusters with the highest deviation degree.
但是，要想充分利用Cognos显示离群值的潜力，需要采用一些更高级的技巧。
To leverage the full potential of Cognos for displaying outliers, however, you need to employ some more advanced features.
一个简单的访问第一节中创建的离群值表BANK . CUSTOMER_OL的查询主题。
A simple query subject that accesses the outlier table BANK.CUSTOMER_OL created in the first section.
在含有离群值的情况下，分析了最小二乘估计不能克服粗大误差对回归曲线的影响。
On the data including the discrete points, least square estimation can not control the strong error affecting the regression curve.
一个是主页面，显示按职业分组的客户记录的一个概述，其中分别列出每种职业中离群值的数量。
The main page that shows an overview of the customer records grouped by profession with the respective number of outliers for each profession.
对于这个报告，Cognos项目中需要两个查询主题，然后通过连接它们获得每个离群值的文本描述。
For your report, you need two query subjects in the Cognos project that you then join to obtain a textual description for each outlier.
接下来将了解InfoSphere Warehouse如何检测离群值，以及如何对数据应用偏差检测。
In the following, learn how InfoSphere Warehouse detects outliers and how you can apply deviation detection to your data.
如此这般的离群值不应被忽略，相反，相当深入研究，以便为帮助我们理解当前和未来事件提供线索。
Outliers such as these should not be ignored but rather studied closely for clues that might help us understand current and future events.
如此这般的离群值不应被忽略，相反，相当深入研究，以便为帮助我们理解当前和未来事件提供线索。
Outliers such as these should not be ignored but rather studied closely for clues that might help us understand current and future events.
应用推荐