为什么基于过滤器的功能选择仅为数字列/功能提供权重？ [英] Why filter based feature selection gives weightage only to numeric columns/features ?

问题描述

Hello Team，

Hello Team,

我的数据集包含数字，字母数字，字符串功能（列）。因此，我试图找出哪些是我可以使用我的数据集来评估模型的最佳功能。我在我的实验
中使用基于过滤器的功能选择模块，并且我运行了我的实验以检查重量，但结果是该模型总是给出数值的更高权重和（0）字符串和字母数字列的零权重。

My data set contains numeric, alphanumeric, string features (columns). So I am trying to figure out what are best features which I can use with my data set to evaluate the model. I am using the filter based feature selection module in my experiment and I ran my experiment to check the weightage but in result this model always gives higher weightage to numeric values and (0) zero weightage to string and alphanumeric columns.

你能否指导我为什么会这样？或者我在这个过程中遗漏了什么？

Could you please guide me why this happening or else I am missing something in this process?

问候，

RPBH

推荐答案

嗨RPBH，

在  基于过滤器的特征选择方法，您需要实现一个算法（度量）到
解释数据集。   什么是基于过滤器的
功能选择以及使用它的原因？段落说明这更详细，并且

示例部分包含三个不同的数据集样本，可通过众多可用指标之一进行评估。

In the Filter Based Feature Selection method, you need to implement an algorithm (metric) to interpret the dataset. The What is filter-based feature selection and why use it? paragraph explains this in greater detail and the examples section contains three differant data set samples, to be evaluated by one of the many metrics available.

你有这个全部设置吗？也许需要进行一些调整？评估哪些列来生成当前分数？

Did you get this all set up? Perhaps some adjustment needs to take place? What are the columns being evaluated to generate the current score?

谢谢，

Mike

这篇关于为什么基于过滤器的功能选择仅为数字列/功能提供权重？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！