训练实验与预测实验 [英] Training experiment vs. Predictive Experiment

问题描述

您好，

我今天花了几个小时尝试配置简单实验，并允许我的简单程序通过Web服务使用它。

I've spent several hours today trying to configure simple experiment, and allow my straightforward program to utilize it by Web Service.

谈到训练实验 - 我实际上完成了一切，就像它在巡演中完成的一样 - 并且每一件工作都很好。

当我尝试重做Predictive Experiment中的所有内容时麻烦开始。我有两种类型的标签"0"。和"1"。每当我将数据传递到实验中时（无论是来自数据集还是来自Web服务，评估
结果都是"1"....

当在Trainig实验中处理相同的数据时....它的工作原理正常....

我尝试过不同类型的模型但我仍然得到了相同的问题。

When it comes to Training Experiment - I've actually done everything like it is done in a tour - and everyting works fine.
Trouble starts when I try to redo everything in Predictive Experiment. I have two types of label "0" and "1". Every time I'm passing the data into the experiment (it doesn't matter if it comes from the Data Set or from Web Service, evaluation outcome is "1"....
And when working on the same piece of data in Trainig experiment.... It works perectly fine....
I've tried with different types of models and still I'm getting the same issue.

任何建议？

推荐答案

Hello Wyplosz，

Hello Wyplosz,

这可能是因为在设置Web服务之前创建预测性实验时删除的模块，或者Web服务输入和输出模块设置不正确。请查看所需的模块是否全部预测实验中是否有
，输入和输出模块设置是否正确？ 

This could be because of the modules you removed while creating the predictive experiment before setting up the web service or the web service input and output modules are not set correctly. Could you please review if the modules that are required are all available in the predictive experiment and if input and output modules are setup correctly? 

您可以完成这三部分  DOCUME ntation
关于预测性实验设置并使用最后一个
教程，检查您的预测实验是否设置在类似的行中。

You can go through this three part documentation about predictive experiment setup and use the last tutorial to check if your predictive experiment is setup in similar lines.

这篇关于训练实验与预测实验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！