search

Alberta-Geological-Survey / depth-to-bedrock

Code repository for the 'Evaluating spatially enabled machine learning approaches for depth to bedrock mapping' PLOS ONE article
MIT License
5 stars 1 forks source link

Geophysics #2

Closed RichardScottOZ closed 1 day ago

RichardScottOZ commented 3 months ago

Hi @stevenpawley,

How do you think you would change this if you were going to add geophysics into the mix at scale?

stevenpawley commented 2 months ago

Hi Richard, would love to be able to put some geophysics into the mix - I think it could help in those deep sediment areas. With the right predictors, you also might not really get any benefit to having those spatial autocorrelation features, which appear to mostly increase the model's accuracy for those deep sediment areas, where the current range of predictors are less relevant to predicting DTB