The invention belongs to the technical field of comment text analysis, and particularly relates to a customer
comment analysis method and
system supporting word frequency statistics. According to the method, each comment is associated with the shop exploration task and the check item identifier, so that the comment data directly corresponds to the business index, an enterprise can conveniently and quickly track a problem source, the
management efficiency is improved, and semantic-related comment contents are grouped by utilizing word segmentation
processing, lexical element feature vectors and
semantic clustering; refined semantic analysis of the comment text is realized, semantic
confusion possibly brought by simple word frequency statistics is effectively avoided, a low-
score check item problem is highlighted in word frequency statistics through performance weighted word frequency calculation, an enterprise can preferentially pay attention to a key business problem, business guidance of an analysis result is enhanced, and the
business efficiency is improved. And a user portrait attribute distribution variance self-adaptive down-regulation mechanism is adopted, so that abnormal word frequency caused by deviation of a specific store or a
user group is effectively inhibited, and an analysis result is more stable and reliable.