I have trained a model to predict the "scout group" based on an structured input text string. eg:
INPUT column NOTE: "Enter ticket holder full name (person attending): John Doe | Enter Group & Section (ie 99th Portsmouth, Cubs): 70th Portsmouth squirrels".
OUTPUT column GROUP: "70th Portsmouth".
This is a table called TRAINING_DATA and the prediction is 100% when adding a new row to this table(!).
The problem I have is moving away from the TRAINING_DATA table and trying to apply the PREDICT("group") in my live data, which is a second table called ORDERS. ORDERS has a NOTE column and a GROUP column. I have variously set the GROUP column to have PREDICT("group") in the Initial Value or the Formula settings for the column. But, nothing happens when I add a new row in the ORDERS table which includes a NOTE.
I had anticipated that the GROUP column in the ORDER table would update - using NOTE against the predictive model. It does not.
How should I approach this?
Here's some context on my use case that might help.
The PAYMENTS and ORDERS tables are used to create the TICKETS table.
I have a BOT triggered by new rows in PAYMENTS to add a new line to TICKETS.
Somewhere in either ORDERS or TICKETS I want to include the PREDICT("group") column - in the screenshot I've put it in TICKETS.
The challenge I have is in getting the PREDICT("Group") column populated. I've tried everything I can think of - putting the PREDICT in Formulas or in Initial Value in either or both tables (ORDERS / TICKETS)... but nothing seems to work. In some instances, if I manually edit the row it will show the prediction.
How can I build the TICKET table, pulling data from PAYMENTS and ORDERS while also adding in the PREDICTed value for Group.
User | Count |
---|---|
15 | |
12 | |
8 | |
8 | |
4 |