Predictive Intelligence , eficient inputs

[sunkarasath]

tell any one for few imp tables in servicenow , when we train the model to predict the which values needs which inputs fields 

for example 

INCIDENT 

1. for category - short_description , description

2. assignment group - short_description , priority , impact 

we dont use use field CI because it has millions of unique values 

so any one worked on PI please give in this format about few tables 

[Its_Azar]

Hi there @sunkarasath 

Use text fields (short_description, description) for NLP and use low/medium cardinality fields (category, priority, impact) and you rightly mentioned not o use high-cardinality fields like cmdb_ci, sys_id. below are some examples.

1. INCIDENT

Predict Category
Inputs:

short_description

description

Predict Subcategory
Inputs:

short_description

description

category

   2. REQUEST ITEM (sc_req_item)

Predict Assignment Group
Inputs:

cat_item

short_description

key catalog variables

requested_for.company

Predict Approval Group
Inputs:

cat_item

variables

requested_for.company

location

3. CASE (sn_customerservice_case)

Predict Case Category
Inputs:

short_description

description

case_type

Predict Assignment Group
Inputs:

short_description

description

case_category

priority

account

4.  CHANGE (change_request)

Predict Assignment Group
Inputs:

short_description

description

change_type

risk

category

5. HR CASE (sn_hr_core_case)
Predict HR Service / Assignment Group
Inputs:

short_description

description

hr_service

employee_type

location