Problem 1: Table 1 shows term frequencies for 3 documents in the collection of 806791 documents. The last column (df) shows term document frequencies. Which two documents are most similar? Use cosine similarity and tf.idf weights? [30 points]

term	Document1	Document2	Document3	df
data	27	4	24	18,165
mining	3	33	0	6723
learning	0	33	29	19241
big	14	0	17	25235

Table 1: Term-Document Matrix