Content Comparison

library(ForestForesight)

# Step 1: Set up the parameters

# Make sure you have DATA_FOLDER in config.yml in your working directory
ff_folder <- Sys.getenv("DATA_FOLDER")
country_code <- "PER"
train_start <- "2023-06-01"
train_end <- "2023-12-31"
validation_start <- "2023-06-01"
validation_end <- "2023-12-01"
prediction_date <- "2024-01-01"

# Step 2: Create date ranges for training and validation
train_dates <- ForestForesight::daterange(train_start, train_end)
validation_dates <- ForestForesight::daterange(validation_start, validation_end)

# Step 3: Set up file paths for saving results
model_save_path <- paste0(ff_folder, "peru_deforestation_model.model")
predictions_save_path <- paste0(ff_folder, "peru_deforestation_prediction.tif")
accuracy_csv_path <- paste0(ff_folder, "peru_deforestation_accuracy.csv")
importance_csv_path <- paste(ff_folder, "peru_feature_importance.csv")

# Step 4: Run the ff_run function

prediction_result <- ff_run(
   country = country_code,
   prediction_dates = prediction_date,
   ff_folder = ff_folder,
   train_dates = train_dates,
   validation_dates = validation_dates,
   model_save_path = model_save_path,
   predictions_save_path_predictions =  predictions_save_path,
   accuracy_output_csvpath =  accuracy_csv_path,
   importance_csvoutput_path = importance_csv_path,
   verbose = TRUE,
   autoscale_sample = TRUE
)

# Step 5: Plot the prediction result
if (!is.na(prediction_result)) {
   plot(prediction_result, main = "Deforestation Prediction for Peru")
}

# Step 6: Print a summary of the results results
ff_cat("Prediction completed. Results saved to disk:")
ff_cat("Model:", model_save_path) 
ff_cat("Prediction raster:", predictions_save_path)
ff_cat("Accuracy CSV:", accuracy_csv_path) 
ff_cat("Feature importance CSV:", importance_csv_path)
# Optional: Load and print the first few lines of the accuracy CSV

if (file.exists(accuracy_csv_path)) {
  accuracy_data <- read.csv(accuracy_csv_path)
  cat("\nFirst few lines of accuracy data:\n")
  print(head(accuracy_data))
}