Review on Geographically Weighted Regression (GWR) approach in spatial analysis

Abstract

<jats:p> In spatial analysis, it is important to identify the nature of the relationship that exists between variables. Normally, it is done by estimating parameters with observations which taken from different spatial units that across a study area where parameters are assumed to be constant across space. However, this is not so as the spatial non-stationarity is a condition in which a simple model cannot explain the relationship between some sets of variables. The nature of the model must alter over space to reflect the structure within the data. Non-stationarity means that the relationship between variables under study varies from one location to another depending on physical factors of the environment that are spatially autocorrelated. Geographically Weighted Regression (GWR) is a technique in which it applied to capture the variation by calibrating a multiple regression model, which allows different relationships to exist at different points in space. A robust algorithm has been successfully used in spatial analysis. GWR can theoretically integrate geographical location, altitude, and other factors for spatial analysis estimations, and reflects the non-stationary spatial relationship between these variables. The main goal of this study is to review the potential of the GWR in modelling the spatial relationship between variables either dependent or independent and its used as the spatial prediction models. Based on the application of GWR such as house property indicates that GWR is the best model in estimating the parameters. Hence, from the GWR model, the significance of the variation can also be tested</jats:p>

The Common Agricultural Policy (CAP) in Extremadura (SW Spain) during the Period 2014–2020:New Opportunities for Economic Diversification in Rural Areas?

“…GWR is a spatial regression technique that provides a local model of the variable and process that makes a forecast by fitting a regression equation to each entity in the dataset, [43,62], in this case, the municipalities of Extremadura. In this way, the individual equations are formed through the incorporation of a dependent variable and several explanatory variables of the entities that are located in each destination entity, in order to obtain the geographical variations [63][64][65]. Thus, the dependent variable is the total amount of CAP aid and the explanatory variables are the percentage of UAA with respect to the total surface area of the municipality, population growth in the period 2014-2020, the standardization of the size of the municipalities into three groups according to the criteria established by the NSI (less than 2000 inhabitants, between 2000 and 10,000 inhabitants and more than 10,000 inhabitants), the unemployment rate, the contracts in the agricultural sector in this period and the Gross Domestic Product (GDP).…”

Mapping geographical inequalities of incomplete immunization in Ethiopia: a spatial with multilevel analysis

“…Considering the spatially varying relationships of variables in clustered data, Ordinary Linear Regression (OLR) and Geographically Weighted Regression (GWR) were considered to identify predictors of incomplete immunization ( 26 , 27 ). However, the assumptions of OLS were not satisfied.…”

Mapping and Assessment of the Spatial and Temporal Distribution of Turbidity in Lake Buhi From Sentinel-2 Images Using Geographically Weighted Regression and Normalized Difference Turbidity Index

“…GWR is a regression technique that extends global regression into local relationships between the independent variable and dependent variables at different locations. The idea of GWR is represented by the model (Sulekan & Syed Jamaludin, 2020):…”