归一化 [英] de-normalize

问题描述

是否可能在训练实验中对数据进行非规范化处理，以查看分数模型中预测列的实际值?

我以前用对数正态化对数据进行归一化使用推荐算法.现在，我看到的值是-0.640057、0.649354但我想看看

I normalize data with lognormal before using recommendation algorithm . Now I see values like -0.640057, 0.649354  but I want to see the real values without normalization.

推荐答案

不幸的是，没有内置模块可以这样做.罗普(Roope)在他的旧线程之一中提出了一些建议:https://social.msdn.microsoft.com/Forums/en-US/ac193e58-ae47-4e8d-8610-d222855a0cc4/data-denormalization? forum =机器学习

Unfortunately, there is no inbuilt module to do so. Roope has some suggestions on how to do it in one of his old threads: https://social.msdn.microsoft.com/Forums/en-US/ac193e58-ae47-4e8d-8610-d222855a0cc4/data-denormalization?forum=MachineLearning

此致，
Jaya

Regards,
Jaya

这篇关于归一化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！