Predictive Intelligence Issue

[Faisal_CG]

Hi all
I am trying to create a classification solution. I have 6858 incidents. I got these many incidents by installing the ITSM demo data plugin. 

This dataset has 105 P1, 121 P2, 1101 P3, 2243 P4 and 3288 P5 incidents.
In terms of state, 36 are New, 92 are In progress, 6 are On Hold, 13 are Resolved and 6711 are Closed.

Categories are as follows: Inquiry Help - 1131, Network 452 and Software 5275.

I gave short description as the input field and Category as the output field and trained the solution on about 5500 records. 

It is showing Precision as 100% and Coverage  and recall as 0.

Why? What should i do?

[Tanushree Maiti]

Hi @Faisal_CG 

1) "10k records is an absolute minimum for training a Classification solution and Servicenow recommends 30k for better prediction results"  . Ensure you have that.

Refer: https://www.servicenow.com/community/itsm-forum/predictive-intelligence-error-while-training-solutio...

Training Task Intelligence Models with Small Data Sets

2)Refer: KB0781894 Predictive Intelligence (PI) FAQ 

"

How many records are needed to train a viable solution?
The recommended number of records for training a good solution is from 30,000 through 300,000. If you submit more than 300,000 records, the most recent 300,000 records are used to train the solution.

13. Can classes be excluded from prediction?

Yes. 

A. Under the Class Confidence record (you want to exclude), select the Precision Coverage Lookup combination record where Precision = 100 and Coverage = 0

B. Select Apply Values from the Actions list. Note, this applies only to classification solutions.

"