we now have code for easily setting up classification pipelines. We should do this for several classifiers, and several hyperparameters for each, to one or more ensembles of results. The result of each could e considered a vote and tallied up for each profile. This could be normalised by the number of results so that the tallies add up to one in a probability like fashion. So for example if 4 classifier vote T4 sippican and 6 classifiers vote T7 sippican, we can say 0.4 prob of T4 and 0.6 prob of T7.
we now have code for easily setting up classification pipelines. We should do this for several classifiers, and several hyperparameters for each, to one or more ensembles of results. The result of each could e considered a vote and tallied up for each profile. This could be normalised by the number of results so that the tallies add up to one in a probability like fashion. So for example if 4 classifier vote T4 sippican and 6 classifiers vote T7 sippican, we can say 0.4 prob of T4 and 0.6 prob of T7.