Aplicación de modelos matemáticos mediante estadística espacial para identificar zonas con alta incidencia en el delito del robo, en la ciudad de Cuenca, Ecuador en el año 2019.

Abstract

The objective was to identify geographical areas with a high incidence of crime, specifically the crime of robbery in its different modalities such as people, homes, economic units and vehicles, registered in the city of Cuenca, Ecuador in 2019, through the application of mathematical models using spatial statistics for area data. The spatial mathematical models chosen were Spatial Delay, Spatial Error and Durbin. These were applied using the R statistical software through its RStudio environment, using the appropriate spatial packages, which allowed the identification of urban census areas with high incidence. In addition, the relationship with socio-demographic variables was analyzed at the same geographical level where these indicators were calculated, in order to relate them to the crime of robbery in the above-mentioned modalities, analyzing also the spatial autocorrelation and the degree of relationship that exists in the variables corresponding to the socio-demographic indicators with the crimes. It was found that, in fact, these present a close relationship with the crimes that occurred in each of the zones of the city; the different spatial models to adequately estimate this type of concentration in the geographic area of study. It was concluded that the best model to estimate this problem is the Spatial Error Model (SEM) for the crimes of robbery of persons and vehicles, the Spatial Autoregressive Model (SAR) for the residences and the Durbin Spatial Model (SDM) for the economic units, since these are the best fit, which was verified by means of the Akaike Statistical Quality Indicator (AIC). It is recommended that researchers or analysts in different fields use this type of spatial model to determine concentrations of this crime in different time periods in each geographical area used in this study.