Check out our Dengue forecasting paper in @pnas.org. We show that ensembling several base models can provide improved forecasts in virtually all countries and locations (more than 180) within country. This is true even if the base models are individually relatively uncalibrated and unsophisticated.

@nunetsi.bsky.social 𝗠𝗮𝘂𝗿𝗶𝗰𝗶𝗼 𝗦𝗮𝗻𝘁𝗶𝗹𝗹𝗮𝗻𝗮 + #MIGHTE_Lab, and 𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗝𝗼𝗵𝗮𝗻𝘀𝘀𝗼𝗻 just published 2 studies on #dengue, a disease that affects nearly 400 million people annually. Their findings could improve forecasting and public health interventions.

PNAS shorturl.at/1BeaN
STM shorturl.at/hPwJA