Improving a tf-idf weighted document vector embedding

Craig W. Schmidt

We examine a number of methods to compute a dense vector embedding for a document in a corpus, given a set of word vectors such as those from word2vec or GloVe. We describe two met…