METHOD FOR BUILDING CHARACTER SEQUENCE DICTIONARY, METHOD FOR SEARCHING CHARACTER SEQUENCE DICTIONARY, AND SYSTEM FOR PROCESSING CHARACTER SEQUENCE DICTIONARY
A multicore CPU (101) of a character sequence data analysis device (1) spreads, across a memory (102), a plurality of blocks obtained by dividing a character sequence dictionary (112), and performs an entry registration process in which unregistered character sequences constituting character sequence data (111) are registered in respective blocks as new entries, one character at a time from the last character of each character sequence, wherein: the entry registration process is performed for a group of blocks in a parallel manner, the blocks in said group of blocks being capable of being processed independently of each other; and when all of the unregistered character sequences have been registered in the respective blocks, a character sequence obtained by concatenating the character sequences registered in the entries of the blocks is output as BW-converted data for the character sequence dictionary (112), in which the character sequence data (111) is already registered.
