In initial stage when I performed decoding, recognition accuracy is low due repetition of words in evaluation speech signal which causes insertion errors. Now, I found some method that automatically corrects the reference transcription if there is repetition. There are some misidentification also. Due to that even though there is no repetition in signal transcription is changed.
When I try to decode using corrected transcription I am getting more deletion and substitution errors and insertion error is reduced. Is there any way to solve this problem. Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In initial stage when I performed decoding, recognition accuracy is low due repetition of words in evaluation speech signal which causes insertion errors. Now, I found some method that automatically corrects the reference transcription if there is repetition. There are some misidentification also. Due to that even though there is no repetition in signal transcription is changed.
When I try to decode using corrected transcription I am getting more deletion and substitution errors and insertion error is reduced. Is there any way to solve this problem. Thanks in advance.