博客
关于我
强烈建议你试试无所不能的chatGPT,快点击我
Predictive Analytics for Business
阅读量:5293 次
发布时间:2019-06-14

本文共 1041 字,大约阅读时间需要 3 分钟。

Business Issue Understanding:

1. What decision need to be made?

2. What information need to inform that decision?

3. What type of ANALYSIS will provide the information to inform that decision?

Uderstanding the Data:

1.What factors would like to drive the result? (time/data/temperature)

2.A good understanding of the data.(Does the data reflect the real situation?)

Preparing the data:

1.Gathering the data: Collecting the data from mutiple sources.

2.Cleaning the data:Deal with incorrect or missing data.

3.Formating the data: Rename the data.

4.Blind the data: combining the data.

5.Sampleing the data.

Analysis/Modeling:

1.Build perdictive model.

2.Vaildate the model.

3.Perdict the future.

4.Repeat the process.

Validate the model:

Check whether the model can perdict the future. If not. iterate the model.

Presentation and visulization:

According to the data, create a stroy of the data to convince audinence.

1. Visualize the data.

2. Reference data source.

3. Measure the success.

转载于:https://www.cnblogs.com/kingoscar/p/6049695.html

你可能感兴趣的文章
内存地址对齐
查看>>
看门狗 (监控芯片)
查看>>
#ifndef #define #endif
查看>>
css背景样式
查看>>
JavaScript介绍
查看>>
js中函数与对象的使用
查看>>
正则表达式
查看>>
开源网络漏洞扫描软件
查看>>
yum 命令跳过特定(指定)软件包升级方法
查看>>
创新课程管理系统数据库设计心得
查看>>
Hallo wolrd!
查看>>
16下学期进度条2
查看>>
前端页面卡顿-代码优化
查看>>
nagios通过脚本对系统进行定制监控
查看>>
jquery判断点击鼠标左、中、右键事件
查看>>
java线程池原理
查看>>
为什么子类的构造方法在运行之前,必须调用父类的构造方法?能不能反过来?为什么不能反过来?...
查看>>
Could not resolve view with name '***' in servlet with name 'dispatcher'
查看>>
springBoot配置elasticsearch搜索
查看>>
Chapter 3 Phenomenon——12
查看>>