In Microsoft ignite 2017, Azure ML team announce new on-premises tools for doing machine learning. this tools much more comprehensive as it provides
1- a workspace helps data wrangling
2- Data Visualization
3-Easy to deploy
4-Support Python codes
in this post and next posts, I will share my experiment with working this tools.
first I got excited right away download it from
after installing I saw below page
So, I have to go through the “Azure Portal” to create an “Experimentation Account”.
so I open the “https://portal.azure.com”.
then I start to create a “Machine Learning Experimentation”
then I need to set up the parameters as below:
however after creation ML experiment you able to add new ML model.
as you see in above picture, we able to download “Azure ML workbench” from there, or we using Mac OS we also able to download it. Moreover, you able to create an account for machine learning model. in below the picture, I created “Machine Learning Model Managment”
then, you should have both in your Azure portal as below:
Now I am able to open the “Azure ML workbench”
So as you see in above there are no Projects, I have to create one. As it is my first time to work with this tool, I am going to try one of the sample projects there. Under example panel, search for iris classification
So, I am going to create something new a below:
just click on the plus sign (number 1), then click on the New Project “number 2”
So, just put a name for the project name, project directory, and then simply create it.
Now, you should see below picture,
this project is a dataset about different types of flowers with their petal and sepal length. this sample is going to run a prediction a group (classification) to predict if we have specific sepal or petal length these flowers belong to which type of flowers.
so for the first process, is to run this project. in this video we able to specify where to deploy the project to the local PC, docker-spark, and docker-python.
Also, for this practice, I am going to run “iris_sklearn.py”. however, before running the project you need to set up some configuration in command prompt
Then, in the file write “pip install matplotlib”, you should see below results after running the code.
Then just write “az login” it will ask you to go to the “http://aka.ms/devicelogin” and provides a code that you should put there.
right after you log in you should see the below message in command prompt.
Now, it is time to just run the code, just now run the code as below
You should see a green sign on the right side of the window. Moreover, in middle page, you will see some more explanation of model run.
in the next posts, I will explain how to explore the results, how to clean data and do the data visualization.