I think the article is more mistitled than anything else. I mean, it's a fine article for someone interested in how to train a classifier on an annotated datas

I think the article is more mistitled than anything else. I mean, it's a fine article for someone interested in how to train a classifier on an annotated dataset in Mathematica. It's just not very interesting for someone interested in how to classify Japanese characters from the Edo period.