Is there a way to check if a string represents an English sentence? I am currently looking at java packages but I cannot find anything that does it yet.
i.e.
the weather is good (valid)
the good is weather (invalid)
Any ideas?
Thanks
Asked
Active
Viewed 3,212 times
5
-
6And how do you define a sentence? – nullpotent Sep 11 '12 at 19:58
-
check if it starts with a capital, ends with a period and contains spaces. – 11684 Sep 11 '12 at 19:59
-
What do you need the sentence for? are you using it as a variable or something? or just to print? are you using a form or some kind of applet? – Gilbert V Sep 11 '12 at 20:02
-
1@pst am I supposed to understand that? – 11684 Sep 11 '12 at 20:02
-
1I mean if it is a valid sentence: i.e. the weather is good (valid) the good is weather (invalid) – user1663930 Sep 11 '12 at 20:02
-
You mean to check sentence structure? – Gilbert V Sep 11 '12 at 20:03
-
1Yeah i want to check the structure. Something with POS tagging – user1663930 Sep 11 '12 at 20:04
-
1POS Tagging. you know that a similar questions was asked here.They have a list of java library's to use. http://stackoverflow.com/questions/2293636/what-is-a-good-java-library-for-parts-of-speech-tagging you can look there for assistance. – Gilbert V Sep 11 '12 at 20:06
-
2I have done the POS Tagging. I need to know if what i get is valid as a sentence – user1663930 Sep 11 '12 at 20:07
-
What POS Tagger did you use? the answer depends on what you use – Gilbert V Sep 11 '12 at 20:09
-
Stanford NLP. But it doesn't check validity – user1663930 Sep 11 '12 at 20:12
2 Answers
1
What you really need is NLP parser (grammatical analysis).
This one is written in Java and it is fantastic.
nullpotent
- 9,162
- 1
- 31
- 42
-
1I have that but i am not sure if it recognizes whether the sentence is valid. It does the tagging – user1663930 Sep 11 '12 at 20:10
0
While I realize the OP referenced Java, I will offer that the MS Office suite has a limited grammar checking engine that is intended to perform very top-level corrections/suggestions for what I would imagine is a fairly simplified set of sentence patterns it recognizes, eg, those found in typical office correspondence. I don't think it pretends to be exhaustive or anything like a full English analysis tool, but it might serve the basic purpose in a limited form.
The check is exposed through COM via the CheckGrammar method, which is called from the Application object of the host Office application. You supply the text to be checked. It can identify several problems in addition to fragments, such as tense and number disagreements, so you might have to refine the results it offers. It's far from perfect, but does give it the old college try :)
David W
- 10,062
- 34
- 60