Leveraging high-resolution weather information to predict hail damage claims: a spatial point process for replicated point patterns (Q2682985)

The authors of the paper adopt a spatial point pattern viewpoint for modeling hail insurance claims. They propose a spatial mixed-effects framework for replicated point patterns to model the frequency and geographical distribution of hail damage claims following a hailstorm. Assume there are \(m\) hailstorms. Let \(\mathbf{Y}_i\) be the spatial point process associated with the \(i\)-th hailstorm, i.e.\( \mathbf{Y}_i\) is a random countable subset of \( S\subseteq\mathbb{R}^2\). The point process \(\mathbf{Y}_i\) can be studied through a counting measure. For a Borel set \(B\in S\), let \(N_i(B) =\#\{\mathbf{Y}_i \cap B\}\) be the count function, i.e. the random number of points falling in \( B\). By supposing that \(\mathbf{Y}_i\) is the Poisson spatial point process with the conditional intensity of special form, authors of the paper derive a series of propositions on jointly distribution for \(N_i(B_1), N_i(B_2),\ldots, N_i(B_k)\) with a Borel sets \(B_1,B_2,\ldots, B_k\). Further, authors discuss the likelihood-based inference and demonstrate estimation of parameters describing conditional intensity. Authors show that incorporating granular data to the proposed model the development of claim reporting patterns helps insurers anticipate and manage claims more efficiently.