You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Mar 24, 2021. It is now read-only.
Rationale:
In general, I'd expect my forecasting biases to be topic dependent and horizon dependent. The horizon is already being recorded (when will you know), it'd be useful to tag predictions.
For example, I may realize that I am being over confident about work related predictions and under confident about leisure related predictions. I also suspect I am less biased for predictions about the near future than predictions about the future.
Right now, the graph displaying the bias is pretty basic. I'd like to be able to download my prediction history as an XML or CSV file so I can look at the data through various angle. This data might also be used to train a machine learning algorithm that debiases predictions based on topic, horizon, etc. If I had enough predictions and cared enough, I could even set up a contest on Kaggle to see who can come up with the best model of my bias.
The text was updated successfully, but these errors were encountered:
I second this. I've been using the site a lot in the last year, and I'd like to go through my data to see what types of predictions I'm good at and what I'm bad at (e.g. I've noticed I'm bad at movie predictions). So I can adjust my calibration.
Both tags, and getting my data out would help a lot with that. If it's too much work to fix this, does anyone have a work around?
We support both of these suggestions.
We're unlikely to get to coding either of them very soon, but would accept pull requests that implemented these changes (if they passed our quality standards).
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Feature request:
Rationale:
In general, I'd expect my forecasting biases to be topic dependent and horizon dependent. The horizon is already being recorded (when will you know), it'd be useful to tag predictions.
For example, I may realize that I am being over confident about work related predictions and under confident about leisure related predictions. I also suspect I am less biased for predictions about the near future than predictions about the future.
Right now, the graph displaying the bias is pretty basic. I'd like to be able to download my prediction history as an XML or CSV file so I can look at the data through various angle. This data might also be used to train a machine learning algorithm that debiases predictions based on topic, horizon, etc. If I had enough predictions and cared enough, I could even set up a contest on Kaggle to see who can come up with the best model of my bias.
The text was updated successfully, but these errors were encountered: