“转换为数据集"的目的是什么?在这个实验中? [英] What is the purpose of "Convert to Dataset" in this experiment ?

问题描述

我正在看下面的实验.

https://gallery.azure.ai/Experiment/Learning-with-Counts-Binary-Classification-2

https://gallery.azure.ai/Experiment/Learning-with-Counts-Binary-Classification-2

如果您查看天气数据集，则某些列的值会很多，例如压力趋势"指的是压力趋势".具有超过300000的缺失值.天气数据集下方是转换为数据集"模块.在属性面板中，设置动作" 到"SetMissingValues"，"Custom Missing Value"设置为"M".如果您查看该模块的输出，则列中的缺失值(例如PressureTendency)不会设置为"M".有什么作用?

If you look at the weather dataset, some columns have a lot of missing values, e.g. "PressureTendency" has over 300000 missing values. Below the weather dataset is the Convert to Dataset module.  In the property panel, "Action' is set to "SetMissingValues", "Custom Missing Value" is set to "M". If you look at the out put of this module, the missing values in columns (e.g. PressureTendency) are not set to "M". What gives ?

预先感谢

弗雷德

---------------------------------

---------------------------------

推荐答案

弗雷德，

对不起，您遇到的麻烦.压力趋势是数字特征. "M"表示是一个字符串值-因此该模块不会替换它，因为它必须将列向下转换为字符串.只有字符串列 替换的值.您可以根据需要使用元数据编辑器将该列转换为字符串，以设置缺少的值.

Sorry for the trouble you are having.  Pressure tendency is a numeric feature.  "M" is a string value - hence the module does not replace it, because it would have to downcast the column to string.  Only string columns have the value replaced.  You can convert the column to string prior to setting the missing values if you prefer by using the metadata editor.

希望这可以帮助并清除模块中的逻辑！

Hope this helps and clears up the logic in the module!

谢谢你，伊利亚

这篇关于“转换为数据集"的目的是什么?在这个实验中?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！