You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@opennlp.apache.org by vinay ds <vi...@gmail.com> on 2011/07/28 09:26:59 UTC
openNLP v/s sharpNLP
I tried POS tag with both c# and java.
But for some words i am getting different results. For example : "melting
point" .
In C# (using sharpNLP model) i am getting result : [NN,NN]
but,
In java, (using openNLP 1.5 model) i am getting result : [VBG,NN] .
Are these kind of difference expected. which one i should follow ?
Thanks in advance,
Vinay
Re: openNLP v/s sharpNLP
Posted by vinay ds <vi...@gmail.com>.
yes. sharpNLP is ported C# version of java.
i used sharpNLP model in c#, and openNLP in java. I cannot use sharpNLP
model in java.
But i am expecting same result .
Thanks
Vinay
On Thu, Jul 28, 2011 at 1:23 PM, Jörn Kottmann <ko...@gmail.com> wrote:
> On 7/28/11 9:29 AM, vinay ds wrote:
>
>> Hi,
>>
>>
>> I tried POS tag with both c# and java.
>>
>> But for some words i am getting different results. For example : "melting
>> point" .
>>
>> In C# (using sharpNLP model) i am getting result : [NN,NN]
>>
>> but,
>>
>> In java, (using openNLP 1.5 model) i am getting result : [VBG,NN] .
>>
>>
>>
>> Are these kind of difference expected. which one i should follow ?
>>
>>
>>
> The OpenNLP POS Model was trained on news texts, you should give it
> an entire sentence to get good results.
>
> I do not know anything about sharpNLP, looks like it is a reimplementation
> of OpenNLP in C#. If they are using our model then it might be a difference
> in the maxent
> or feature generation code.
>
> Do you need to use OpenNLP from C#?
>
> Jörn
>
Re: openNLP v/s sharpNLP
Posted by Jörn Kottmann <ko...@gmail.com>.
On 7/28/11 9:29 AM, vinay ds wrote:
> Hi,
>
>
> I tried POS tag with both c# and java.
>
> But for some words i am getting different results. For example : "melting
> point" .
>
> In C# (using sharpNLP model) i am getting result : [NN,NN]
>
> but,
>
> In java, (using openNLP 1.5 model) i am getting result : [VBG,NN] .
>
>
>
> Are these kind of difference expected. which one i should follow ?
>
>
The OpenNLP POS Model was trained on news texts, you should give it
an entire sentence to get good results.
I do not know anything about sharpNLP, looks like it is a reimplementation
of OpenNLP in C#. If they are using our model then it might be a
difference in the maxent
or feature generation code.
Do you need to use OpenNLP from C#?
Jörn
openNLP v/s sharpNLP
Posted by vinay ds <vi...@gmail.com>.
Hi,
I tried POS tag with both c# and java.
But for some words i am getting different results. For example : "melting
point" .
In C# (using sharpNLP model) i am getting result : [NN,NN]
but,
In java, (using openNLP 1.5 model) i am getting result : [VBG,NN] .
Are these kind of difference expected. which one i should follow ?
Thanks in advance,
Vinay
Re: openNLP v/s sharpNLP
Posted by Jörn Kottmann <ko...@gmail.com>.
You can also try to use IKVM, as described here in our wiki:
https://cwiki.apache.org/OPENNLP/a-quick-guide-to-using-opennlp-from-net.html
Jörn
On 7/28/11 9:26 AM, vinay ds wrote:
> I tried POS tag with both c# and java.
>
> But for some words i am getting different results. For example : "melting
> point" .
>
> In C# (using sharpNLP model) i am getting result : [NN,NN]
>
> but,
>
> In java, (using openNLP 1.5 model) i am getting result : [VBG,NN] .
>
>
>
> Are these kind of difference expected. which one i should follow ?
>
>
>
> Thanks in advance,
>
> Vinay
>