underdarkGIS’s Twitter Archive—№ 10,436

RT @DABaranger: Linear regressions with 2nd/3rd order polynomials were equal to or better than neural-nets (aka 'deep learning') in every d…