EpiInvert

EpiInvert analyzes and forecasts epidemiological time series by decomposing incidence curves, comparing related time series, and estimating the daily reproduction number R t of infectious diseases.

Key Features:

• EpiInvert: Inputs raw daily incidence time series and the pandemic time serial interval, and outputs weekly seasonality, an incidence trend, and its reproduction number.
• EpiIndicators: Inputs two related epidemiological time series such as daily incidence and death count, and outputs a daily ratio and delay between both time series.
• EpiInvertForecast: Inputs an incidence trend obtained by EpiInvert and a database of past observed time series, and forecasts the incidence in the next four weeks using the most similar past series.
• Variational techniques: Incorporated into the analytical framework for time series processing.

Scientific Applications:

• Infectious disease surveillance: Estimates the incidence trend and the daily reproduction number R t of any infectious disease from regional time series data.
• Time series comparison: Compares related epidemiological time series such as incidence and death, or incidence and hospital or ICU admission.
• Short-term epidemiological forecasting: Provides four-week incidence forecasts based on similarity to past observed time series.

Methodology:

EpiInvert uses variational techniques for incidence curve decomposition. EpiInvertForecast identifies the most similar past observed time series from a database to generate forecasts.