I was wondering if the following strategy is possible:
First, we use the location of already known ancient cities as well as the geographic information at that location ( climate, terrain) to train a predictive model, which can tell the likelihood of a city being there given a location. Then we can use the likelihood as a priori to augment the prediction of the gravity model.
I was wondering if the following strategy is possible:
First, we use the location of already known ancient cities as well as the geographic information at that location ( climate, terrain) to train a predictive model, which can tell the likelihood of a city being there given a location. Then we can use the likelihood as a priori to augment the prediction of the gravity model.