Understanding Confidence Levels: Criteria for Color Icons

what is the criteria by which you determine the green, orange, and red icons? It maps to confidence level, but anything else? If so, how do you evaluate confidence?

6 comments

Channeled
APP
·
·
The provided information doesn't contain details about green, orange, and red icons or their specific criteria. However, we do have information about confidence evaluation in data scoring systems. For mobile number verification, our system uses weighted scores like P1 (high confidence - 25% connection rate) and P2 (lower confidence - 4-10% connection rate) based on proprietary algorithms that evaluate multiple data signals. Which specific feature or area of Clay are you seeing these colored icons in?
Source: Data Test Methodology | Documentation | Clay University
Richard X.
·
·
I included an image - can you not see it? If no, I'm referencing the Claygent-produced URL finds. The column gives me a "Response" and beside "Response" there's a green dot, orange triangle, or red square. These seem to indicate the confidence level of Claygent's findings, and I'm asking how does evaluate confidence? Do these symbols indicate anything else?
Channeled
APP
·
·
The available information doesn't contain specific details about how Claygent evaluates confidence for the green, orange, and red icons in URL finds or what criteria determine these confidence indicators. What specific type of URLs is Claygent finding for you, and are you noticing any patterns in when certain colored icons appear?
Owen C.
·
·
Hey Richard, thanks for reaching out! Appreciate your patience. : ) Yes the Green, Orange, and Red circles on AI columns are confidence levels. Green is high, yellow is medium, red is low. Typically Claygent will also include it's reasoning as a response within the nested data if you're interested in hearing a bit more about how it came to this conclusion! Let me know if there's anything else we can help with! :)
Richard X.
·
·
thanks Owen C.; the reason I asked is bc Clay came back with a lot of red squares. I plan on using those outputs for outbound msgs, but if a lot of them are "low confidence", should I even be using them? Worst case is I generate a message with erroneous data and damage first impressions.
Owen C.
·
·
Makes total sense! This depends on what the ask is to your AI column. As in if you're asking the AI column to do research, then write the message, the confidence may be related to the data is scraping rather than the quality of the writing. Do you mind sending the link (url) to the table so we can take a look?
(52 kB)

Understanding Confidence Levels: Criteria for Color Icons | Clay