I have to create a web based spelling checker for Hindi Language. I have ~3,00,000 hindi words
Spelling checker will perform these two tasks:
1. Check whether the word is correct or not
2. If it is not correct then provide suggestions
I am creating a web service which will do this task. Various clients like Wordpress Plugins will be developed to consume this service.
Is this the best approach? Or I should do something else?
It is easy to check the word is correct or not. But to provide suggestion it looks like I have to perform many queries for each word.
The algorithm that I thought for spelling suggestions is:
create a list of various "Matras" and "Alphabets"
add each "Matra" and "Alphabet" to the wrong word then check its entry in database.
if found then add it to array of suggested words.
after completing this loop return array of suggested words.
But there are many "Matras" and "Alphabets" in Hindi language. Therefore this will make a lot of database queries. This is the main problem.
Is there any alternative way to get spelling suggestions?
View Complete Post