Prepare your data

The Machine Learning Workbench allows users to build Machine Learning and Deep Learning models offline, using streamed data from devices.

The first step is deciding what data you are looking to use to build your machine learning model.

Basic Configuration

In the Rayven Workflow 

1. Connect the Repository node to your data source (in image below you can see the repository node is connected to a HTTP end point)
2. Define the retention time of data, IE last 30 days, last 60 days
3. Select the data attributes that come in through the Json payload that you wish to upload to the machine learning work bench for training purposes 
   
   
   
   
   
   
4. Go to the Machine learning and add a new workflow
   
   
   
   
   
5. You will see in the left panel under repositories, the repository you created in the work flow.
6. Drag The repository onto the canvas and click on the expand icon (marked in red)
   
   
    
7. In the repo node you will be able to see all the data sent from the workflow to the machine learning bench for training.
   
   
   
   
8. Now you are ready to pre-process your data, for data cleansing and data transformation.
9. Next step go to Actions (on the left side panel) and drag the Pre-processing Node right under the repository node. A blue icon will light up under the repository node, and at that point release the mouse and the pre-processing node will automatically connect to the repository node
   
   
   
   
   
10.  Expand the pre-processing node, and now you are ready to start cleansing your data using the pre built list of pre-processing functions
    
    
    
11. Once you have finished preparing your data, you are now ready to choose a data model.