A method of correcting provisional boundaries of "bunsetsu"

In order to solve the problem that the amount of computer memories required for translating the non-segmented "kana" sentences into the "kanji-kana" sentences grows rapidly in proportion to the increasing of the length of the sentence, a new method of finding provisional boundaries of "bunsetsu" using 2nd-order Markov chain probability has been developed. This paper proposes a method to correct the provisional boundaries of "bunsetsu" for non-segmented "kana" sentences, by looking up all the word candidates in the dictionary. The improvements of "relevance factor" P and "recall factor" R for provisional boundaries of "bunsetsu" determined and corrected by these methods were evaluated by experiment using the statistical data for 70 issues of a daily Japanese newspaper.<
>