FEATURE TEXT STRING-BASED SENSITIVE TEXT DETECTING METHOD AND APPARATUS

A feature text string of a currently detected text is acquired, and the feature text string is detected according to a finite-state machine established in advance, so as to obtain frequency of occurrence of each keyword in the feature text string. A weight of the keyword category in the text is calculated for each keyword category of multiple keyword categories based on the frequency of occurrence of each keyword corresponding to the keyword category and a preset weight of each keyword. The text is determined to be sensitive when the weight of at least one keyword category is greater than a preset threshold.