EVIDENCE SYNTHESIS | DENGUE FORECASTING ARCHITECTURES

AI PhD Committee Presentation Dashboard

WHY THIS PROBLEM REMAINS UNSOLVED

TWO MATHEMATICAL BARRIERS

1

Non-Stationary Stochastic Process

Disease incidence is a multiscale, non-Markovian signal with high-frequency meteorological forcing, seasonal vector population fluctuations, and low-frequency climate regime shifts superimposed.

Fixed-order temporal kernels alias across these bands and distort the phase-space geometry of P(yt | *).

2

Ill-Posed Inverse Operator

The mapping F : c → d violates Hadamard well-posedness:

  • Existence outbreaks without deterministic climate precursors
  • Uniqueness degenerate trajectories from distinct climate states
  • Stability small perturbations trigger nonlinear regime shifts

The Jacobian condition number diverges under covariate multicollinearity, yielding unstable parameter estimation.

PRIOR APPROACH LANDSCAPE (Analyzing... papers)

APPROACH WHAT IT DOES FAILS AT PAPERS

THE ARCHITECTURAL GAP

No existing architecture jointly addresses non-stationarity and ill-posedness while producing calibrated distributional forecasts across geographies.

SYSTEMATIC LITERATURE METRICS (n=0)

Approach Distribution

GRADE Evidence Scoring

*Scoring algorithm based on rigorous study design and relevance to mathematical barriers (Ill-posedness & Non-stationarity).

Barrier Failure Map

GRADE Evidence Database

Year Authors Title/Summary Approach Category GRADE Score