This program implements several machine learning models for mapping income distribution using satellite imagery.


There are several optional command line arguments:

  • --model: Machine Learning algorithm, including 'lr', 'rf', 'knn', 'svm', and 'cnn'.
  • --arch: ResNet architecture, such as 'resnet18' or 'resnet34'.
  • --regularize: Regularization techniques, such as 'ridge' or 'lasso'.
  • --batch-size: Size of a training batch for CNN.
  • --lr: Learning rate for CNN.
  • --epochs: Number of training epochs for CNN.
Share this project: