Geo-Profiling: beyond the Current Limits. A Preliminary Study of Mathematical Methods to Improve the Monitoring of Invasive Species

Abstract

The Geographic Profiling (GP) is a data analysis tool that has great potential. Presently, it is used only minimally, and is almost always used "as it is", independently on other analysis or data processing methods. GP was initially created as a forensic tool, to find the origin of a series of events (crimes) done by a single actor. However, using this method in integration with others, it is possible to enlarge the opportunities of geographical data analysis. The promising results of this method in integration with others, even if some of them are quite well known methods since many years-and thus well tested-show a number of further possible applications. Here we treat data clustering and partitioning with Kmeans and Dbscan methods; space partitioning (Voronoi tessellation) and a method to assign weights to the events constituting the data set.

Spatially clustered count data provide more efficient search strategies in invasion biology and disease control

“…Here, the effect of time was not explored, rather its units were set to the time interval in which data were collected. As suggested in many studies (Rossmo 2000, Raine et al 2009, Santosuosso and Papini 2018, a more accurate geographic profiling model is one that considers temporal variability in the data to draw its inferences.…”

Utilization of Herbaria in Ecological Studies: Biodiversity and Landscape Monitoring

“…Geographic profiling (GP) (with the "Rossmo formula") was effective in assessing the spread of invasive Caulerpa in the Mediterranean region (Papini et al, 2013). This suggests that GP is a more practical search strategy in identifying the source populations of invasive species; fueled by datasets, predictive modeling enables scientists to anticipate species responses to climate change and other environmental shifts (Santosuosso & Papini, 2018;Stevenson et al, 2012).…”

Tracing outliers in the dataset of Drosophila suzukii records with the Isolation Forest method

“…Also the DBSCAN [10,36] clustering method is able to identify outliers but requires a priori knowledge about the maximum distance for which points on a map are considered adjacent and the minimum number of adjacent points that are to be considered for a cluster. In both cases, it is necessary to find a biological justification for these assumptions.…”

“…This analysis is useful, since it can suggest control methods and give an idea about the gateway of the invasion [8]. Recently, further refinements of the method allowed to improve the power and reliability of PGP, in particular by allowing (a) to clarify if a distribution pattern is caused by more waves of invasion, rather than from a single starting point [8], (b) to evaluate the robustness of the results with a jackknife procedure [26], (c) to give different weights to data on a quantitative basis (on the basis of the population dimension) or new methods of data partitioning [10,36]. The PGP was applied on the DS distribution in Europe, since it does not require any a priori knowledge about the invasion routes, while the biological justifications of the used parameters can be evaluated also by testing several parameters values.…”