基于计数的特征选择如何工作? [英] How does count based feature selection work ?

问题描述

我在理解计数学习"时遇到了麻烦《 Azure ML文档》一章.

I have trouble understanding the "Learning with count" chapter in Azure ML documentation.

https://docs.microsoft.com/zh-cn/azure/machine-learning/studio-module-reference/data-transformation-learning-with-counts

https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/data-transformation-learning-with-counts

如何获取基于计数的特征的样本表"?从下面的屏幕快照中的表格开始吗?

How do you get the "sample table of count based features" from the table in the screenshot below it ?

感谢一百万，

弗雷德

推荐答案

也许您已经可视化了在对每个值(A和B)的类成员进行计数时出现的矩阵.它的格式应为:

maybe you have already visualized the matrix that arises when you count the class membership of each value (A and B). It should have the form:

要编写基于计数的功能表，请将对每个值A和B的成员计数的2列添加到类0和1(在示例中，这些列分别为0_0_Class000_Count和0_0_Class001_Count)，然后查看中的条目 输入值列(即A或B)将这两个新列中的对应计数值放入.例如，在原始表的第一行中，输入值=A.因此，在类0的新列中，第一行中，您必须输入次数 其中原始表中的A = 0，即2，并且在类1的新列中(仍在第一行中)，您必须写出A = 1的次数，表示1.对于行2和3，因为在输入值"列中重复了值A.计数 从第4行开始更改，其中输入值变为B.要容纳的值是上面小矩阵中汇总的值.

To write the table of count-based features, add the 2 columns that count the membership of each value A and B to classes 0 and 1 (in the example these columns are 0_0_Class000_Count and 0_0_Class001_Count, respectively), and looking at the entries in the Input value column (that is A or B) put in these 2 new columns the corresponding counting values. For instance, in the first row of the original table, Input value = A. So in the new column for class 0, in the first row, you have to put the number of times in which A = 0 in the original table, that is 2, and in the new column for class 1, still in the first row, you must write the number of times A =1, it means 1. The same for rows 2 and 3, because in the Input value column the value A is repeated. The counts change from row 4 on, in which Input value becomes B. The values to accomodate are those summarized in the small matrix above.

希望这会有所帮助;)

这篇关于基于计数的特征选择如何工作?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！