The prompting method of IPTV input method for Chinese with full text search technology

In order to prompt the efficiency of the server based Chinese input method for IPTV, the behaviors of querying for the program´s text information are analyzed. Then it´s proposed to integrate the full-text search engine of Sphinx with the input method to mine the accurate associating characters or words. To fit the main querying behavior for the program role´s names, the program synopses are mined for the role names which are the remained continuous character sequences after getting rid of the normal words with the maximum matching segmentation and inverse maximum matching segmentation algorithms. Then a special name library composed of these role names is used for the input method´s associating computation. The order of the associating words is optimized through referring to the program´s demanding ratios which are searched by Sphinx with the inverted index on the name library.