Improve the recognition rate using Chinese characteristics.
l
ling long
started a topic
about 1 year ago
I'm a Chinese and native language is Chinese. Chinese is different from english.
A word in English is a meaning, and there is no use of group words like Chinese.
for example
"I eat apples today" translated into Chinese,
“我今天吃苹果了”
in this example, two Chinese words "今天" means today, "苹果" means apple.
so in Chinese when write an "苹", and the next are most likely to be "果".
and my question is in sdk is there are any interface let me use this feature to impove recognition accuracy.
back to this example, when sdk recognize "苹", and we can recognize "苹果", even "果" is write carelessly.
so in Chinese there are many group words.
Thank you.
1 Comment
O
Olivier @MyScript
said
about 1 year ago
Dear Ling Long,
Currently, I am not sur to understand your questions...
Indeed, our chinese language has been developped for a while, and we never got complaints regarding it. We also worked a lot on language models, so such behavior is not likely to occur... Did you notice some? If so, can you please provid with examples?
Also, which recognizer are you using? The Text, Diagram? Did you set properly the Chinese language?
ling long
I'm a Chinese and native language is Chinese. Chinese is different from english.
A word in English is a meaning, and there is no use of group words like Chinese.
for example
"I eat apples today" translated into Chinese,
“我今天吃苹果了”
in this example, two Chinese words "今天" means today, "苹果" means apple.
so in Chinese when write an "苹", and the next are most likely to be "果".
and my question is in sdk is there are any interface let me use this feature to impove recognition accuracy.
back to this example, when sdk recognize "苹", and we can recognize "苹果", even "果" is write carelessly.
so in Chinese there are many group words.
Thank you.