A new method of error correction using 2nd-order Markov models of forward, backward and middle type in Japanese syllables

2nd-order Markov model in Japanese is useful to correct errors in "kanji-kana" "bunsetsu" and reducing the ambiguity of speech recognition candidates (syllable) in "bunsetsu", but not to detect and correct errors located at boundaries of "bunsetsu" in a sentence. This paper proposes a new method to correct such erroneous syllables using three types of 2nd-order Markov model, forward, backward and middle types. In an experiment, it is found that H/sub 2f/=H/sub 2b//spl ges/H/sub 2m/, where H/sub 2f/, H/sub 2b/ and H/sub 2m/ denotes the entropy of forward, backward and middle type respectively. Corresponding to the location of an erroneous syllable in bunsetsu, the accumulative accuracy rate, which denotes the rate of correct syllables included in 1st candidate or up to 10th candidates is presented as follows: 1. in the case of the top location in bunsetsu (type S/sub 1/): C/sub 2b//spl ges/C/sub 2m//spl ges/C/sub 2f/; 2. in the case of the second location in bunsetsu (type S/sub 2/): C/sub 2m//spl ges/C/sub 2f//spl ges/C/sub 2b/; 3. in other cases (type S/sub 3/): C/sub 2f//spl ges/C/sub 2m//spl ges/C/sub 2b/ where C/sub 2f/, C/sub 2b/ or C/sub 2m/ is accuracy rate by using forward, backward or middle Markov models respectively. From this result, it is found that the error correction capability is improved by applying backward type to S/sub 1/, middle type to S/sub 2/, and forward type to S/sub 3/.