How to use Azure's Machine Learning learning modules differently.

Asked 2 months ago, Updated 2 months ago, 4 views

I'm an office worker who started studying machine learning last week and is not familiar with technology.
Sorry for the basic question

I started studying to predict the number of visitors at Azure's Machine Learning.
I'm stuck because I don't know how to use the learning module

Even if it's the same linear regression
How do I use Liner Regression and Bayesian Liner Regression?
(Do I have to check the accuracy every time I can't use it differently...?)

When creating logic like this for other modules, use this module
If you know or have a summary site,
Please let me know

Azure ML Learning Module List

machine-learning azure

2022-09-30 13:46

3 Answers

Microsoft has officially released a cheat sheet!

https://docs.microsoft.com/ja-jp/azure/machine-learning/studio/algorithm-cheat-sheet


2022-09-30 13:46

I've just started machine learning myself, but if you're studying machine learning individually, I recommend scikit-learn, which is developed by many volunteers.For scikit-learn, Head office manual is well-equipped, and if you are not good at English, there is a reference book written in Japanese.There is a lot of information on the web.

For scikit-learn, how to use the modules differently is illustrated in page of your home documentation:

Enter a description of the image here

Also, it is quite difficult to build an environment for machine learning, but Google colab allows you to use the Jupiter Notebook with a machine learning module for free, so you can start learning right away.


2022-09-30 13:46

I think there are two ways to deal with it.

  • Use multiple learning algorithms in parallel to verify accuracy without considering using them differently
  • Understand individual learning algorithms, check the status of the data you want to use, and select the algorithm

In the former case, you will compare the metrics calculated by the Evaluate Model module on Azure Machine Learning Studio to see which learning algorithms had better accuracy.

As for the latter, we will proceed with the study in books and sites that mention the theory of machine learning, but the content of the data used depends on which algorithm is appropriate for what we want to achieve.Therefore, even if you make the same customer forecast twice, if the data is different on the first and second occasions (e.g., whether or not to include the number of people passing in front of the store), the appropriate learning algorithm may be different.


2022-09-30 13:46

If you have any answers or tips


© 2022 OneMinuteCode. All rights reserved.