无法弄清楚我应该对数据应用哪种算法 [英] Unable to figure out which algorithm should i apply to my data

问题描述

你好团队

我具有以下数据模式，我想应用分类算法来了解客户是否打算在特定月份或特定时期接受订单，但是数据的问题是它不包含任何直接的分类/因变量，这将有助于 我要应用两类分类算法.

I have below pattern of data and i want apply classification algorithm to know whether customer going to accept order in particular month or period, but the problem with data is it dose not contain any direct categorical/dependent variable which will help me to apply two class classification algorithm.

那么您能建议我任何想法或想法吗，我该如何准备数据集?还是我需要尝试其他算法才能达到目标?

So could you please suggest me any idea or thought around it that how can i prepare dataset? or do I need to try different algorithm to achieve my target.

希望您能理解我的问题陈述，作为我的初学者，请帮助我.

I hope you understand my problem statement, please help me as I am beginner in this.

此致

RPBH

 

推荐答案

论坛中的其他用户可能还有其他建议.

Other users in the forum might have additional suggestions.

这是我的建议:您需要进行功能设计，才能从这些现有的原始功能中创建其他功能.示例-创建一个功能来统计最近1/2/3个等月内接受的订单数，总订单数 您可能需要尝试其他功能，领域专家可能会帮助您确定与用例相关的其他功能.生成要素后，您将需要根据历史记录创建标签. 数据，请检查以确保在为分类模型定义标签时没有标签泄漏.

Here is my suggestion: you would need to do feature engineering to create additional features from these existing raw features. Example - create a feature which counts number of accepted orders over last 1/2/3 etc months, count of number of total orders for the customer etc. You would need to try different features and perhaps a domain expert can help you with what additional features might be relevant in your use case. Once you have the features generated, you would then need to create labels based on historic data, pls check to ensure that there is no label leakage when defining the label for the classification model. 

此致，
Jaya

Regards,
Jaya

这篇关于无法弄清楚我应该对数据应用哪种算法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！