How to easily use ANN for prediction mapping using GIS data?

Artificial Neural Networks (ANNs) are a powerful component of Artificial Intelligence (AI), finding applications across diverse fields like social sciences, medicine, and applied engineering. ANNs are known for their high reliability and validity, further enhanced by flexible configuration options. When combined with spatial data, ANNs significantly improve the confidence in results, often outperforming traditional regression or classification techniques, particularly in prediction mapping applications.

This comprehensive course provides a step-by-step guide to understanding and implementing ANNs using the NeuralNet Package in R and real-world landslide data and thematic maps. You'll gain practical skills with readily available data, code, and materials.

Here's a detailed breakdown of what you'll learn:

• Data Preparation: Learn to produce training and testing datasets using automated tools in QGIS (or utilize your own datasets).
• Neural Network Modeling:
  • Run the Neural Net function with training and testing data.
  • Visualize the NN function network.
  • Analyze pairwise NN model results of explanatory and response data.
  • Examine Generalized Weights plots for explanatory and response data.
  • Determine variable importance using the NNet Package function.
• Advanced Analysis:
  • Run the NNet function for sensitivity analysis.
  • Perform sensitivity analysis of explanatory and response data.
• Prediction & Validation:
  • Run the Neural Net function for prediction with validation data.
  • Evaluate prediction validation results using AUC value and ROC plot.
  • Produce prediction maps using raster data.
• Thematic Map Processing:
  • Import and process thematic maps, including resampling, stacking, and categorical to numeric conversion.
• Exporting Results:
  • Export the final prediction map as a raster.tif file.

Important Resource: LaGriSU Version 2023_03_09 (Free) is available for download via GitHub (search for /Althuwaynee/LaGriSU_Landslide-Grid-and-Slope-Units-QGIS_ToolPack). LaGriSU offers automatic extraction of training/testing thematic data using Grid and Slope units, streamlining your workflow.