#co-occurrence-matrix

Question 1: Select all that apply

Three words have these rows in a term–term co-occurrence matrix (counts with the context words data, fashion and model):

Which statements are true? Select all that apply.

Co-occurrence rows are long and sparse, while word2vec vectors are short and dense1
Passing the co-occurrence matrix to truncated SVD (LSA) gives short, dense word vectors2
word2vec needs no manually labelled training data3
Doing poorly on word2vec's prediction task is fine, since only the learned weights matter4
By dot product, word 1 is more similar to word 2 than to word 35