I'm working on a project of tagging articles by keywords, topics, etc.
I've searched the web for related articles and info, but didn't find anything that is actually relevant. I currently thought of doing it in collecting form somewhere a tagged data-set of articles, use word2vec on it, and run it through a cnn as we did in lesson 5, when the output will be a list of around 1K tags, as in imagenet. do you think this is a good idea?
will be glad to get any info about such tasks, especially relevant data sets.