The GeoAI toolbox contains tools for using and training AI models that work with geospatial and tabular data. These tools use modern machine learning and deep learning techniques and integrate them with GIS.
The tools in the GeoAI toolbox allow you to train and use models that perform classification and regression on feature and tabular datasets, as well as classify, transform, and extract information from unstructured text using natural language processing (NLP).
The tools in the Feature and Tabular Analysis toolset use automated machine learning to train, fine-tune, and ensemble the best machine learning models given the data and available compute resources. The trained models can be used for predicting both categorical variables (classification) and continuous variables (regression) on similar datasets. The tools in the Text Analysis toolset allow you to use and fine-tune pretrained text and NLP models from ArcGIS Living Atlas of the World, or create models given labelled text data. The tools in this toolset also work with models created using the ArcGIS API for Python arcgis.learn module. The models created by these tools can be used in, or further fine-tuned using, ArcGIS API for Python.
All tools in the GeoAI toolbox require the installation of the required deep learning frameworks libraries. For instructions on installing deep learning packages, see Deep Learning Libraries Installers for ArcGIS.
Shapefiles cannot store null values. Tools or other procedures that create shapefiles from nonshapefile inputs may store or interpret null values as zero. In some cases, null values are stored as very large negative values in shapefiles, which can lead to unexpected results. See Geoprocessing considerations for shapefile output for more information.
The Feature and Tabular Analysis toolset contains tools for applying machine learning and deep learning algorithms to feature or tabular data.
The Text Analysis toolset contains tools that perform natural language processing on text. Text can be classified or transformed, and entities such as addresses can be extracted.