In this paper, we present a semi-supervised clustering method for microblog in which both word-level and microblog (document)-level constraints are automatically generated totally based on statistical information rather than any kind of external know
