Retail Range Prediction

rangeml.txt

ML Lifecycle for Range Prediction in Retail

Step 1: Model Creation 
MindsDB lets you connect to many databases, data warehouses, data lakes, etc., via the CREATE DATABASE command. 
       
       #Query for Database Creation
       CREATE DATABASE minds_range_db      --- display name for database. 
        WITH ENGINE = 'postgres',           --- name of the mindsdb handler 
        PARAMETERS = { 
            "user": "postgres",             --- Your database user. 
            "password": "postgres",         --- Your password. 
            "host": "0.tcp.in.ngrok.io",    --- host, it can be an ip or an url. 
            "port": "14424",                --- common port is 5432. 
            "database": "range_minds_db"    --- The name of your database *optional. 
        }; 

Step 2: Training 
Once the database is connected and we have data in our AI Tables we can create a model. 
The CREATE MODEL statement creates and trains a machine learning (ML) model. Below is the syntax to create a regression model: 

        #Query for Model Creation & Training       
        CREATE MODEL mindsdb.range_dates_predictor FROM minds_range_db (SELECT * FROM product_info)    
        PREDICT on_range_number, off_range_number; 

Step 3: Monitor Training 
Once the model query is run, it takes some time for the model to get trained based on the size of dataset. One will be able to check the status of the model using the below query:            
        
        #Query for Model Training Monitor       
        SELECT status,error FROM mindsdb.models WHERE name='range_dates_predictor'; 
   
Step 4: QA Test with limited set of products 
Once the predictor model status sets to complete, we are now ready to make predictions. 
Making predictions is now nothing but querying the predictor model as if it were a table. 
The SELECT statement lets you make predictions, follow the below query: 

        #Query for Model Testing       
        SELECT on_range_date,on_range_date_explain,off_range_date,off_range_date_explain FROM mindsdb.range_dates_predictor   
        WHERE primary_color = 'black' AND secondary_color ='blue' AND family_name = 'JACKETS' AND class_name = 'BOMBER'
        
Step 5: Batch Predict/Serve 
After testing and playing out with the predictor now we ready to make batch predictions with real datasets by joining multiple tables. 

        #Query for Model Prediction
        SELECT pi.on_range_date AS 'real_on_range_date',m.on_range_date AS 'predicted_on_range_date', 
              pi.off_range_date AS 'real_off_range_date', m.off_range_date AS 'predicted_off_range_date',
              pi.primary_color,pi.secondary_color,pi.family_name,pi.class_name 
        FROM product_info AS pi JOIN  mindsdb.range_dates_predictor AS m