{"id":"arxiv/2611.13590","url":"https://arxiv.org/abs/2611.13590","tags":["math.DS","q-bio.PE","stat.AP"],"score":0,"title":"Fractional-Order Hybrid SIR Forecasting of COVID-19 in Bangladesh and the United States: Memory Effects and Out-of-Sample Evaluation","domain":"math.DS","source":"arxiv","authors":["Md Babul Hossain","Mohammad Shah Alam"],"reviews":0,"subject":"math.DS","summary":"Fractional-order epidemic models are often justified by memory, yet whether memory improves forecasts is rarely tested out of sample. We extend a hybrid SIR framework, in which a daily transmission rate is reconstructed from reported cases and then forecast, to Caputo fractional order $α\\in(0,1]$. We prove well-posedness, conservation, positivity and a linearized threshold at disease-free states, and derive a Grünwald-Letnikov reconstruction that reduces exactly to the integer-order one at $α=1$. Our main theoretical result shows that for $α<1$ the influence of the unknown initial infectious…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-11-10","reactions":{}}