[smc-discuss] Project Idea: Parse Date expressions in Malayalam

Santhosh Thottingal santhosh.thottingal at gmail.com
Thu Oct 14 04:20:57 PDT 2010





---- On Thu, 14 Oct 2010 03:53:43 -0700 Praveen A  wrote ---- 

>Santhosh,
> Give a use case of this analysis. I would think integrating it email clients makes sense. Like you can add an appointment to calendar without an actual calendar invite.

Ok, this got some importance in semantic analysis of Malayalam language. 
It is still a challenge for us to develop a system which can analyse a sentence and generate a meaningful unambigous data representation out of that. Looking at the sentence, can we say who is 'karma', 'kriya' etc? No. That kind of language analysis can only be achieved through solving lots of smal small problems. The question is how much a computer can understand from a malayalam sentence by parsing it.  Consider this date parsing as a small step towards that.

Praveen is correct. You should have noticed that gmail extracts the date information out of email content and asks whether a calendar event shold be created from that or not. Try this by composing an email with some mention about date and time. Will it work for all languages? Don't we need it?

Consider we want to search for some information that happened on some specific date from a big data collection in malayalam. eg: all news articles which talks about december 6 in Malayalam.

Actually I put that idea in list only in academic interest. I invite the members of this list to find and propose interesting use cases of this.


-Santhosh





More information about the discuss mailing list