Janus visualization and settings on UTA Level 1
Janus is a powerful extension of SmartSimple’s Grants, Research and Case Management Systems that enables organizations of all types to look at their past to help predict the future.
With Janus, you can easily identify which funding applications have the greatest chances of success, or which medical claims are most likely to be fraudulent.
Organizations can select as many inputs as needed, along with a single output. Janus then determines how these inputs affect the output. Once Janus has been trained, these results - and the likelihood of the outcomes - are applied to any new applications or cases submitted within your SmartSimple system.
1. Go to UTA Settings - General tab and click on Janus Machine Learning link.
2. Click on New Model tab.
3. You will be presented the New Model screen, including the following fields:
- Name - free text field to enter appropriate name
- Variable - read only field
- Description - free text field to enter appropriate description
- Algorithm - drop-down to select the Janus Algorithms to be used
- Outcome - drop-down to select the field that will reflect outcome
- Use for Prediction - toggle to indicate if model is used for prediction
- Model Recordset - select filter for subset of records in UTA to be used in Model
- Model Building Progress - read-only field that displays progress once model submitted to model builder
- Last Modified - read-only field showing date of last modification
- Modified By - read-only field showing name of person who made change
- Last Build Error - read-only field showing any build errors
- Model Validation - indicates if model validation is correct, based on whether there is an outcome enabled
4. Populate fields as appropriate and click Save.
5. Additional buttons are exposed along the bottom of the screen as follows:
- Submit to Model Builder
The goal of Janus is to help organizations determine those applications that have the greatest chances of success, or which ones are most likely to have issues.
However, to do this they must select as many fields as needed to act as inputs and also can select only a single output.
Janus then determines how these inputs affect the output.
The steps for adding the fields as as follows:
1. Click on the Fields button at the bottom of the Janus model.
2. The resulting screen has a fields for configuring New Attributes as well as a list of fields under the Model Attributes header. These listed fields are those custom fields (with sets of pre-determined field types) found in records returned by the selected filter in the AI model.
3. The New Attributes section includes the following options:
- Order - value increases in increments of 10, but can be manually overwritten
- Is Outcome - indicate if the selected field is the outcome. All Janus models must have single outcome.
- Field - drop-down showing certain fields available to the entities returned by the selected filter in the AI model
4. Select the field, update the order if needed, indicate if its an outcome of needed, then click on the Add Attribute button.
5. The field will now be listed in the Model Attributes section.
6. Click on the Remove button in the Model Attributes section for any of the fields you wish to remove.
Users must be in a role that has permission to access the Janus tab.
Go to the UTA Settings - Security tab, access the Feature and Function Permissions section.
Click on the Level 1 tab, go to the Show Janus Tab section and update the Logic or Role options as required.
If a user has appropriate permissions then they will see the Janus tab when they access a level one in a UTA.
If they click on the Janus tab then they will the likelihood of each possible Outcome option. The likelihood is based on an algorithm that uses the selected fields.