From f20be4fbb91236e3a802c90cc486f18e601dd2dd Mon Sep 17 00:00:00 2001 From: Nikos Bosse <37978797+nikosbosse@users.noreply.github.com> Date: Tue, 3 Dec 2024 12:21:24 +0000 Subject: [PATCH] Update Metadata to explain interpolation. --- model-metadata/Metaculus-cp.yaml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/model-metadata/Metaculus-cp.yaml b/model-metadata/Metaculus-cp.yaml index 3c98a9ad..338a31f8 100644 --- a/model-metadata/Metaculus-cp.yaml +++ b/model-metadata/Metaculus-cp.yaml @@ -21,8 +21,8 @@ repo_url: "https://github.com/Metaculus/respiratory-diseases" license: "CC-BY-4.0" designated_model: true team_funding: "This project is supported by the National Science Foundation under Award No. 2438211. Any opinions, findings and conclusions or recommendations expressed in this project are those of Metaculus and our forecasters, and do not necessarily reflect the views of the National Science Foundation." -methods: "A recency-weighted average of predictions made by forecasters on the Metaculus prediction platform." -data_inputs: "Users are allowed to make use of any data they choose. The recency-weighted average takes only the numeric forecasts made by forecasters on the platform into account." +methods: "A recency-weighted average of predictions made by forecasters on the Metaculus prediction platform. Missing forecasts are linearly interpolated." +data_inputs: "Users are allowed to make use of any data they choose. The recency-weighted average takes only the numeric forecasts made by forecasters on the platform into account. We only launch new questions every 2 weeks. This means that alternatingly, only the forecasts 1, 3, 5 or 0, 2, 4 will actually be available. Missing forecasts are linearly interpolated to always produce a set for horizons 1, 2, 3, 4." methods_long: "The Metaculus Community Prediction is a consensus of recent forecaster predictions. It is designed to respond to big changes in forecaster opinion while still being fairly insensitive to outliers. For every forecaster, on ly their most recent prediction is kept. Predictions are assigned a number n, from oldest to newest (oldest is 1). Every prediction is weighted proportional to exp(sqrt(n)). Predictions are then aggregated by creating a mixture distribution of all available weighted forecasts." ensemble_of_models: true ensemble_of_hub_models: false