Spatial regression is a particular type of regression that introduces space or geographical context into the statistical framework of regression for prediction or inferring causal relationships.

Although spatial regression can be done to some extent through traditional regression algorithms based on spatial feature engineering, its main task and focus is to provide unique regression algorithms that consider spatial relationships such as spatial dependence or spatial heterogeneity. Some example use cases include predicting house prices based on census data and location information, or finding a home considering the property’s proximity to economic opportunities, schools, health care, and roadways for commutes.

The following spatial regression algorithms are supported:

• Spatial Cross-Regressive Model (SLX)
• Spatial Lag Model (SAR)
• Spatial Error Model (SEM)
• Geographical Regressor (GR)
• Geographically Weighted Regression (GWR)
• Spatial Regimes (OLS_Regimes)
• Spatial Fixed Effects (SFE)