- This topic has 3 replies, 3 voices, and was last updated 9 months, 1 week ago by
.
-
Posts
-
This year things are a little different. First, the submission format is a JSON file as opposed to CSV. Second, there is much more latitude in how you submit your answers due to the nature of the PDF. If you want to use another PDF type please let us know.
Because of these differences, we are open to having a webinar to discuss the process. Early feedback is appreciated.
Dear staff
Suppose we can directly estimate the torque margin. Why must we estimate the torque margin expressed at a probability distribution function (PDF)? Thanks.Hi Hanqisu,
We are further pursuing methods of explainability this year. Last year, we had participants submit a discreet distribution with a confidence metric. This year, we are building on that to ask for a continuous distribution (which is required for a regression analysis). As indicated in the writeup, the scoring algorithm evaluates the probability of your estimate at the exact torque margin.
If you like, we can incorporate a delta function, which will allow you to submit an exact value. However, if this value is off even by a little, you’ll get a score of 0 because you’ll have predicted a probability of zero. If you consider the problem, there are better ways to submit a PDF that will give you a better chance (and get closer to what you’re trying to achieve).
- You must be logged in to reply to this topic.