Generate Confidence Models for an Existing Predictive Model
If you want to generate confidence models for an existing predictive model, you do not need to retrain the model. You can create confidence models via the services on the Validation Thing. First you will need to Retrieve a Predictive Model URI to use as a parameter. Then follow the steps below.
1. In ThingWorx Composer, open the Validation Thing and navigate to the Services tab.
2. Open the CreateJob service.
3. Enter the following minimum parameters:
|
Parameter
|
Description
|
||
|---|---|---|---|
|
jobName
|
A name to reference the job.
|
||
|
datasetRef
|
Click to open a dialog box, select Add, and enter a datasetUri and a format (must be parquet).
|
||
|
modelUri
|
Enter the ID of an existing predictive model. Use the following format: results:/models/<model ID>
|
||
|
goalField
|
Field name of the goal variable.
|
||
|
createConfidenceModel
|
Click True to enable confidence model creation.
|
||
|
confidenceLevels
|
Enter a number greater than 0 and less than 1 that will be used to calculate the likelihood of an actual value falling within a specific range. The default value is 0.8.
|
4. Click Execute.
A validation job runs on the existing predictive model. A new PMML model is generated with the confidence model information appended to the existing information. When the job completes, a jobID is returned in the Output panel.
5. Copy the jobID. If you want to run predictive scoring with all of the confidence model information, follow the Retrieve the Confidence Model URI procedure for the new model and use the resulting model ID in a predictive scoring job. The predictive scoring results will include columns for the confidence model you generated.