Question

1
Replies
455
Views
 Hero Tolosa (HeroT143)
Accenture

Accenture
AE
HeroT143 Member since 2018 2 posts
Accenture
Posted: September 26, 2018
Last activity: September 27, 2018
Posted: 26 Sep 2018 5:07 EDT
Last activity: 27 Sep 2018 8:40 EDT
Closed

Predictive Analytics Output Discussion

I need to understand the attached file for decision making and building predictive models for post estimation. Please give some reference or discussion regarding to my concern.

Thanks.

Decision Management System Administration
Moderation Team has archived post, This thread is closed to future replies. Content and links will no longer be updated. If you have the same/similar Question, please write a new Question.
Posted: 2 years ago
Posted: 27 Sep 2018 8:40 EDT
Ivar Siccama (Ivar_Siccama)
PEGA
Senior Product Manager - Predictive Analytics
Pegasystems Inc.
NL

Hi HeroT143,


The screenshot is the output of running a predictive model.


Result: this is the label associated to the Class. The predictive model rule allows you to set your own label. In strategies this will be available as pxResult. 


Score: this the output score of the predictive model


Class: this is the score interval number. For example, if your score distribution was divided into deciles, this would be the index of the decile


Percentage: this is the number of cases (expressed in a percentage) of the population that are assigned to this Class (score interval)


Behavior: this is the behavior related to this Class. The meaning depends on the outcome you are predicting. For example if the outcome is the defaulting on a loan then the behavior would be the probability to default


Lift: this the ratio of behavior in this Class to the overall behavior


These outputs are also available to use in a Strategy. When you use the predictive model in a strategy, check out the 'Output mapping' tab,


Regards,


Ivar


Let's add a picture from https://mesh.pega.com/docs/DOC-90494

Show More

Hi HeroT143,


The screenshot is the output of running a predictive model.


Result: this is the label associated to the Class. The predictive model rule allows you to set your own label. In strategies this will be available as pxResult. 


Score: this the output score of the predictive model


Class: this is the score interval number. For example, if your score distribution was divided into deciles, this would be the index of the decile


Percentage: this is the number of cases (expressed in a percentage) of the population that are assigned to this Class (score interval)


Behavior: this is the behavior related to this Class. The meaning depends on the outcome you are predicting. For example if the outcome is the defaulting on a loan then the behavior would be the probability to default


Lift: this the ratio of behavior in this Class to the overall behavior


These outputs are also available to use in a Strategy. When you use the predictive model in a strategy, check out the 'Output mapping' tab,


Regards,


Ivar


Let's add a picture from https://mesh.pega.com/docs/DOC-90494



 


 



Show Less