Spatial Machine Learning Concepts and Algorithms
Learn about the Spatial AI machine learning concepts.
Spatial Machine Learning is machine learning on spatial and location data to gather usable information and intelligence for various applications.
The majority of data are associated with location. Location and location relationships affect business and event outcomes. This kind of impact of location is called a spatial effect. There are two types of spatial effects, namely, Spatial Dependence and Spatial Heterogeneity. Spatial machine learning incorporates these spatial effects by taking into account the location data (in addition to business or attribute data) to improve the analytical and predictive models.
One approach of spatial machine learning is to use traditional machine learning models by integrating spatially explicit independent variables generated using spatial feature engineering operations. Oracle Spatial AI provides some Spatial Feature Engineering algorithms for this purpose. However, this approach may be limited by what spatially engineered features are used, as these may not be able to fully cover the spatial effects in most application scenarios.
Besides using traditional machine learning models involving spatial feature engineering, Spatial AI allows you to create specialized machine learning models designed to fully incorporate spatial information and spatial relationships or spatial effects (such as spatial dependence and spatial heterogeneity).
In summary, the following table lists the spatial machine learning techniques and the corresponding algorithms supported by Spatial AI. Using these techniques, you can build spatial machine learning models or generate spatially engineered features for use in traditional machine learning models to detect patterns and make predictions.
| Techniques | Algorithms |
|---|---|
| Spatial Feature Engineering | |
| Spatial Clustering | |
| Spatial Anomaly Detection |
|
| Spatial Regression |
|
|
Spatial Classification |