Create your first ML Pipeline

Let’s look at how to create a machine-learning pipeline using the QuickML Platform.

Note : QuickML uses Zoho IAM as authentication manager. In order to use the service, the user must be signed in using Zoho accounts.

Resources: Click here to access some sample datasets to get started with.

Creating ML Pipeline using pipeline builder interface

1. Go to datasets section
2. Click Import Dataset to upload a dataset to QuickML platform using the available data connector options.
   
3. Select the data source from the listed options. The data can be from internal, external, or local file system.
   After uploading the file from the above given options, the data quality score will be displayed. If the data needs to be processed further, it can be done using the below given steps.
4. If the user wants to process the uploaded data, preprocessing can be initiated with Go to Data Cleaning option provided and the pipeline builder interface will be opened to do the required processing steps after giving data pipeline name. Every step in the data pipeline will result in a preview and profile view of processed data for that respective stage.
5. Once finished with the data cleaning, you can continue building machine-learning pipeline by using the Create Pipeline option from Pipelinessection. There you need to choose the dataset uploaded in the above step.
   After creating the complete pipeline using the pipeline builder, the machine learning pipeline can be executed using the Execute option available in editor.

Your first pipeline will be executed, and the pipeline stats will be displayed. The pipeline execution metrics will help you analyze execution time and resource utilization. The screenshot below provides a sample pipeline, and real-time pipelines can be constructed based on specific requirements.