The Evidence Tree provides detail on how Rainbird reached a result as well as detail on where Rainbird sourced the information used to reach the judgment.
Each Evidence Tree is unique and the Evidence Tree for a particular result can be accessed by clicking the information icon that appears next to a result.
Figure 1: A result with a certainty factor
Clicking the information icon will open a new tab which displays the Evidence Tree for that result:
Figure 2: Example evidence tree
The Evidence Tree is a visual representation of how Rainbird made a decision, broken down into fact cards about the various rules, relationships, and conditions Rainbird used to inform its decision.
In the example above, a Company, ‘ACME’ , ‘receives final decision’ about a business loan with the Decision ‘Accept’. The certainty factor of 85% has been reached by running the four conditions –
- ACME receives minimum requirements Accept
- ACME’S Credit Score Outcome: Medium
- Product selected: Credit Card
- ACME Outcome based on Registration Date: Long
Clicking on a condition will open up a further fact card where Rainbird will detail evidence about how that condition was satisfied, providing the user with more information about how and why Rainbird made a judgement.
For the Condition ‘Credit Score Outcome a certainty factor of 60% was assigned to a Medium Credit Score. To investigate the fact further, the user can click on a condition to open up another fact card and can follow Rainbird’s decision making process in returning the fact Medium:
To return Credit Score: Medium, Rainbird made a calculation, comparing the Credit Score received against different number brackets. The evidence tree shows that a Credit Score of Medium was achieved by checking whether the inputted score of 713 is between the number bracket of 300 to 1000.
There are different ways to enter or produce facts in Rainbird. The knowledge map will display the fact in a coloured box depending on how that fact was created.
Facts can be created when:
- The end-user has created facts at run time by interacting with the knowledge map via a user interface (shown in red)
- Rainbird access information stored in the knowledge map (shown in orange)
- Rainbird has derived facts from rules in the knowledge map or (shown in blue),
- Imported from a datasource (shown in green, not shown in this example).
For more information on the Evidence Tree, click here to watch a Rainbird training video.
The RBLang below will generate the map used in the example above. Click on ‘Export .rbird’ to download the knowledge map, or ‘copy RBLang’ and paste the code directly into Rainbird.
Query and Results
To see a detailed evidence tree, please run the query on the ‘receives final decision’ relationship.
