Combining multiple probability predictions using a simple logit model

B-Tier
Journal: International Journal of Forecasting
Year: 2014
Volume: 30
Issue: 2
Pages: 344-356

Authors (6)

Satopää, Ville A. (not in RePEc) Baron, Jonathan (not in RePEc) Foster, Dean P. (University of Pennsylvania) Mellers, Barbara A. (not in RePEc) Tetlock, Philip E. (not in RePEc) Ungar, Lyle H. (not in RePEc)

Score contribution per author:

0.335 = (α=2.01 / 6 authors) × 1.0x B-tier

α: calibrated so average coauthorship-adjusted count equals average raw count

Abstract

This paper begins by presenting a simple model of the way in which experts estimate probabilities. The model is then used to construct a likelihood-based aggregation formula for combining multiple probability forecasts. The resulting aggregator has a simple analytical form that depends on a single, easily-interpretable parameter. This makes it computationally simple, attractive for further development, and robust against overfitting. Based on a large-scale dataset in which over 1300 experts tried to predict 69 geopolitical events, our aggregator is found to be superior to several widely-used aggregation algorithms.

Technical Details

RePEc Handle
repec:eee:intfor:v:30:y:2014:i:2:p:344-356
Journal Field
Econometrics
Author Count
6
Added to Database
2026-01-25