Spatial Regression Models for Demographic Analysis

O principal objetivo dessa pesquisa é investigar o quão sensível é a geração de inovação dos estados brasileiros às influências das externalidades de diversificação e de especialização industrial, no período compreendido entre 2001-2011. Além disso, estudar o comportamento da inovação através do território permite a inferência de como políticas públicas de fomento à ciência e tecnologia tem agido no Brasil. Outros fatores determinantes da inovação são também considerados, como a capacidade de investimento em ciência e tecnologia dos estados, o nível de escolaridade e a defasagem temporal da inovação. Como medida de geração de inovação, são utilizados os depósitos de patentes per capita, a fim de mensurar a capacidade tecnológica de cada estado. A base de dados utilizada neste trabalho consiste na conjugação de dados provenientes do Instituto Brasileiro de Geografia e Estatística (IBGE). dos dados de depósitos de patentes do Instituto Nacional de Propriedade Industrial (INPI), de dados sobre gastos com ciência e tecnologia do Ministério da Ciência, Tecnologia e Inovação (MCTI) e dados sobre capital humano do Ministério da Educação (MEC). A metodologia utilizada aborda a Análise Exploratória de Dados Espaciais (AEDE) e modelos de regressão espacial com dados em painel. Tais procedimentos permitiram acompanhar a trajetória da inovação através do território no período em análise e foram estudados os principais marcos legais da inovação brasileira a fim de se estabelecer se esses esforços contribuíram para o fortalecimento da política regional de inovação.

Modeling the impact of land use/land cover (LULC) factors on diurnal and nocturnal Urban Heat Island (UHI) intensities using spatial regression models

Malaria is one of the most severe public health problems worldwide with 300 to 500 million cases and about one million deaths reported to date of which 90% were from world health organization (WHO) Sub Saharan Africa (SSA) countries. The purpose of this study was to explore the spatial distribution of malaria parasite prevalence (MPP) among districts of Southern Nations Nationalities and Peoples Regional State (SNNRS) in Ethiopia by using 2011 malaria indicator survey (MIS) data collected for 76 districts and to model its relationship with different covariates. Exploratory spatial data analysis (ESDA) was conducted followed by implementation of spatial lag model (SLM) and spatial error model (SEM) in GeoDa software. Queen contiguity second order type of spatial weight matrix was applied in order to formalize spatial interaction among districts. From ESDA, we found positive spatial autocorrelation in malaria prevalence rate. Hot spot areas for MPP were found in the eastern and southeast parts of the region. Relying on specification diagnostics and measures of fit, SLM was found to be the best model for explaining the geographical variation of MPP. SLM analysis demonstrated that proportion of households living in earth/local dung plastered floor house, proportion of households living under thatched roof house, average number of rooms/person in a given district, proportion of households who used anti-malaria spray in the last 12 months before the survey, percentage household using mosquito nets and average number of mosquito nets/person in a given district have positive and statistically significant effect on spatial distribution of MPP across districts of SNNPRS. Percentage of households living without access to radio and television has negative and statistically significant effect on spatial distribution of MPP across districts of MPP. Malaria is spatially clustered in space. The implication of the spatial clustering is that, in cases where the decisions on how to allocate funds for interventions needs to have spatial dimension.

Tree height–diameter relationships are usually studied using linear or nonlinear models, but exogenous variables, especially spatially autocorrelated and dependent variables of tree diameter or height, are not often considered in height–diameter modeling. Three types of spatial regression models — spatial lag model, spatial error model, and spatial Durbin process model — are explored in this study. The height–diameter relationships are modeled using the spatial regression models to investigate the effects of spatial dependence and spatial autocorrelation and the roles of the exogenous variables generated by neighboring trees. Case study 1 shows that the spatial lag model should be used to analyze height–diameter relationships, in which heights of neighboring trees, which are exogenous variables, and the endogenous variable DBH significantly affect height growth. Case study 2 shows that the spatial error model performs better than other models, and that height growth is not only affected by its endogenous variable diameter but also by unobserved variables that vary spatially and result in residual spatial autocorrelation. Spatial regression models are an approach to height–diameter modeling that provide insight into how the endogenous variable diameter, the exogenous variables height and (or) diameter of neighboring trees, and locally varied but unobserved environmental or ecological variables contribute to height growth.

The influence of building or dwelling energy performance on the real estate market dynamics and pricing processes is deeply explored, due to the fact that energy efficiency improvement is one of the fundamental reasons for retrofitting the existing housing stock. Nevertheless, the joint effect produced by the building energy performance and the architectural, typological, and physical-technical attributes seems poorly studied. Thus, the aim of this work is to investigate the influence of both energy performance and diverse features on property prices, by performing spatial analyses on a sample of housing properties listed on Turin’s real estate market and on different sub-samples. In particular, Exploratory Spatial Data Analyses (ESDA) statistics, standard hedonic price models (Ordinary Least Squares—OLS) and Spatial Error Models (SEM) are firstly applied on the whole data sample, and then on three different sub-samples: two territorial clusters and a sub-sample representative of the most energy inefficient buildings constructed between 1946 and 1990. Results demonstrate that Energy Performance Certificate (EPC) labels are gaining power in influencing price variations, contrary to the empirical evidence that emerged in some previous studies. Furthermore, the presence of the spatial effects reveals that the impact of energy attributes changes in different sub-markets and thus has to be spatially analysed.

When employing the spatial diagnostics examined in Chapter 4, social scientists will often find evidence of spatial autocorrelation (see, e.g., Eff 2004). As discussed in Chapter 1, this predisposition of social science data toward spatial autocorrelation often results from interdependence between the units studied by social scientists. In other cases, social science data exhibit spatial dependence not as a result of behavioral interdependence but as a consequence of spatial clustering in the sources of behaviors of interest to social scientists. The spatial dependence, in short, may be consistent with either a spatial lag model or a spatial error model. Substantive theory will often lead scholars to believe that a spatial lag specification or a spatial error specification is more appropriate for their particular substantive application. Scholars may, for example, expect that a spatial diffusion process is at work and thus believe that a spatial lag model is warranted. Although such a specification may seem appropriate, such a theoretical expectation should not go untested. It would be inappropriate to estimate a diffusion model with a spatially lagged dependent variable if the spatial dependence diagnosed via, for example, the univariate Moran's I , is instead produced by spatial clustering in the sources of otherwise independent behaviors. This model misspecification will lead the researcher to inappropriate substantive inferences about the nature of the spatial dependence in her data. Inappropriate spatial model specification is all the more problematic because of the close mathematical relationship between a spatial lag model and a spatial error model with spatial autoregressive error dependence. As this chapter will discuss, a spatial autoregressive error model can be rewritten as a spatial Durbin model with both spatially lagged dependent and independent variables if a set of nonlinear common factor constraints are valid. Because of this close relationship between spatial autoregressive dependence in a spatial lag model and spatial autoregressive dependence in a spatial error model, a significant spatial parameter in a spatial lag model may reflect spatial clustering in omitted sources of the behavior of interest rather than true spatial lag dependence consistent with a diffusion process.

Changwu County is a typical soil and water loss area on the Loess Plateau. Soil erosion is an important ecological process, and the impact of land use/cover change on soil erosion has received much attention. The present study used remote sensing images of the study area in 1987, 1997, 2007, and 2017 to analyze the land use/cover change (LULCC), and the RUSLE model was applied to estimate the soil erosion in different times. We exploited the Sankey diagram to visualize the spatiotemporal changes in land use/cover and soil erosion. We planned to obtain the most suitable model by comparing the application of different spatial regression models (Geographically weighted regression model, Spatial lag model, Spatial error model) and Ordinary least squares in LULCC and soil erosion changes. The results revealed that land use/cover has significantly changed in the last 30 years. From 1987 to 1997, cropland expansion came mainly from planted land and orchards, which transformed 68.99 km2 and 64.93 km2, respectively. In 1997–2007, the planted land increase was mainly through the conversion of cropland. In 2007–2017, the increase in orchard area came mainly from cropland. The forest land increase was mainly from the planted land. Soil erosion in Changwu County was dominated by slight erosion and light erosion, although the area of slight erosion and light erosion continued to decrease. The annual average soil erosion increased, which was estimated at 977.84 ton km−2 year−1, 1305.17 ton km−2 year−1, 1310.60 ton km−2 year−1, and 1891.46 ton km−2 year−1 in 1987, 1997, 2007, and 2017, respectively. These amounts of transformation mainly occurred when slight erosion was converted to light erosion, light erosion was converted to moderate erosion, and moderate erosion was converted to light and severe erosion. The Spatial lag model and Spatial error model have higher accuracy than the Geographically weighted regression model and Ordinary least squares when fitting the effect of LULCC and soil erosion change, where the accuracy exceeded 0.62 in different periods.

An empirical evaluation of spatial regression models

In the present study, we investigate different factors related to demography, land use, transportation and height that influence the generation of the PM2.5 contaminant due to woodburning in the conurbation of Temuco and Padre Las Casas, Chile. First, a global regression model was implemented to explore the relationships between the variables, and subsequently, spatial regression models were implemented (geographically weighted regression and spatial error model). Results suggest that height and infrastructure density have a strong association with the PM2.5 contaminant using the global regression and spatial error models. However, results indicate that residential density, distance to high vehicle congestion, and length of local roads within a vicinity of 100 m are significant when employing the geographically weighted regression. In addition, the latter type of regression model presents the best performance when compared to the other two models (global regression and spatial error). These results may help authorities in the allocation of resources to improve air quality in the studied conurbation. Future research should include other types of spatial regression models and other relevant variables that may influence the generation of PM2.5 due to woodburning.

This research discussed about model of Human Development Index (HDI) in Central Java with spatial regression analysis. and identify variables that give significant influence. First, analyze the influence factors based on result of p-value from t test in multiple linear regression models. Then, made spatial weight matrix with queen continguity method. After that, estimate spatial regression models, namely spatial autoregressive (SAR), Spatial error models (SEM), and spatial autoregive moving average (SARMA) and choose the best model based on minimum AIC value. The results showed that SAR was the best spatial regression model and the significant variables was the gross enrollment rates at senior high schools, the health workers, and the district minimum wages. All of them that give positive influences. The variable that give biggest influence for HDI was the health workers.
 Full Article

The purpose of this study was to investigate the determinants of hotel RevPAR using data from 274 hotels in Korea's capital region. As the outset, it is imperative to clarify the concept of location due to fact that there is lack of consensus defining the hotel location. Based on the Mono-Centric Model and the Agglomeration Model, this study proposes an operational definition of hotel location. In addition, traditional explanatory variables such as room count, auxiliary facility size, and brand were considered as determinants. The least squares regression model (OLS), the spatial autoregressive model (SAR), and the spatial error model (SEM) were established in order using cross-sectional data. The SEM is chosen as the final model for this study because it overcomes the problem of spatial autocorrelation. The findings confirm that attributes such as room count, auxiliary facility size, brand, and location are important drivers of RevPAR. The number of companies in the hotel area, in particular, may improve RevPAR, whereas agglomeration may be detrimental to profit in the hotel room market. This study, in addition to having managerial implications, contributes to future research by providing a methodological guide on spatial regression models. It is suggested to spatial econometrics model, which overcoming the problem of spatial autocorrelation, is more appropriate than the traditional OLS in explaining hotel performance.

Land surface temperature (LST) is the instantaneous radiative skin temperature of land obtained from satellite sensors. Measured by visible, infrared or microwave sensors, the LST is useful in determining thermal comfort for urban planning. It also serves as a precursor to many underlying impacts such as health, climate change and the likelihood of rainfall. Due to the paucity of observed data because of cloud cover or rain-bearing clouds in the case of microwave sensors, it is necessary to model LST for the purpose of forecasting. Two spatial regression models were employed: the spatial lag model and the spatial error model. Using Landsat 8 and Shuttle Radar Topography Mission (SRTM), these models can be studied and compared in terms of their robustness in reproducing LST. Whereas LST is to be the independent variable, built-up area, water surface, albedo, elevation, and vegetation are to be considered as dependent variables and their relative contributions to LST examined.•Modeling LST based on spatial regression models with calculated LST as independent variable.•Dependent variables to be considered are normalised difference Built-up index (NDBI), normalised difference vegetation index (NDVI), modified normalised difference water index (MNDWI), albedo and elevation.•The models were validated using k-fold cross validation method, mean square error and standard deviation.

Is It a Curse or a Blessing to Live Near Rich Neighbors? Spatial Analysis and Spillover Effects of House Prices in Beijing

The COVID-19 pandemic began at the end of 2019 in.Wuhan, China, spreading rapidly to all parts of the world. This virus was detected in Indonesia in February 2020, and then the government began to make social restriction policies in March 2020. All countries in the world experienced a fairly severe economic shock. The limited mobility of people impacts decreasing all forms of activities that require face-to-face contacts, such as tourism, transportation, hospitality, etc. It causes a decrease in economic growth which can be measured by the size of the gross regional domestic product. This analysis uses Gross Regional Domestic Product (GRDP) data as a Y variable, general allocation funds (X1), profit-sharing funds (X2), open unemployment rate (X3), labor force participation rate (X4), and the number of industries (X5). It will be analyzed using spatial regression. There are several spatial regressions such as Spatial Autoregressive (SAR) and Spatial Error Model (SEM), Spatial Durbin Model (SDM) and Spatial Durbin Error Model (SDEM). The SDM and SDEM are special cases of SAR and SEM. SDM involves spatial effects on independent and dependent variables, while SDEM involves spatial effects on dependent variable and error. So, we will analyzed using SDM and SDEM. From the results, SDEM is the best method used for the data, and all variables used are significant. The predicted GRDP values of districts/cities in East Java are classified into five classes. There are eight regencies/cities included in the class with declining GRDP due to the COVID-19 pandemic, namely Ngawi Regency, Bangkalan Regency, Pamekasan Regency, Blitar City, Probolinggo City, Pasuruan City, Mojokerto City, and Batu City.

In recent years, accurate estimation and spatial mapping of above-ground carbon (AGC) storage in forests have been crucial for formulating carbon trading policies and promoting sustainable development strategies. Forest structure complexities mean that during their growth, trees may be affected by the surrounding environment, giving rise to spatial autocorrelation and heterogeneity in nearby forest segments. When estimating forest AGC through remote sensing, data saturation can arise in dense forest stands, adding to the uncertainties in AGC estimation. Our study used field-measured stand factors data from 138 forest fire risk plots located in Fenglin County in the Northeastern region, set within a series of temperate forest environments in 2021 and Sentinel-2 remote sensing image data with a spatial resolution of 10 m. Using ordinary least squares (OLS) as a baseline, we constructed and compared it against four spatial regression models, spatial lag model (SLM), spatial error model (SEM), spatial Durbin model (SDM), and geographically weighted regression (GWR), to better understand forest AGC spatial distribution. The results of local spatial analysis reveal significant spatial effects among plot data. The GWR model outperformed others with an R2 value of 0.695 and the lowest rRMSE at 0.273, considering spatial heterogeneity and extending the threshold range for AGC estimation. To address the challenge of light saturation during AGC estimation, we deployed traditional linear functions, the generalized additive model (GAM), and the quantile generalized additive model (QGAM). AGC light saturation values derived from QGAM most accurately reflect the actual conditions, with the forests in Fenglin County exhibiting a light saturation range of 108.832 to 129.894 Mg/ha. The GWR effectively alleviated the impact of data saturation, thereby reducing the uncertainty of AGC spatial distribution in Fenglin County. Overall, accurate predictions of large-scale forest carbon storage provide valuable guidance for forest management, forest conservation, and the promotion of sustainable development strategies.