Hi
I definitely used the "en-pos-perceptron.bin" in both cases.
However when switch to the "en-pos-maxent.bin" with 1.5.2 I do get
results that I would expect. So it seams the described issue only
affects the perceptron model.
Here the detailed results for the maxent model.
Sentence:
A nice travel to the biggest volcano of Mexico.
POSTaggerME#topKSequences(tokens) with openNLP 1.5.2 and "en-pos-maxent.bin"
-0.19092493831060522 [DT, JJ, NN, TO, DT, JJS, NN, IN, NNP, .]
Detailed Probabilities:
[0.9574457453640682, 0.988073661574357, 0.9962495810301972,
0.9709960280930423, 0.978828016790525, 0.9730651680770073,
0.9857216847007988, 0.9886882554359132, 0.9889536036052374,
0.9834568956230735]
On Tue, May 22, 2012 at 4:46 PM, Jörn Kottmann <[email protected]> wrote:
> On 05/22/2012 04:34 PM, Jeyendran Balakrishnan wrote:
>>
>> Hi Rupert, are you sure you are loading a model from
>> "en--pos-perceptron.bin", but passing it to a POSTaggerME tagger?
>> Unless it is a typo in your email, that could be the problem (ie passing a
>> perceptron model to a maxent classifier)...
>
>
> OpenNLP is auto detecting the model which is passed in, but that would
> easily explain
> why the probs are different, e.g using maxent first and then the perceptron
> model.
> But he mentioned in an earlier mail that he used the *same* model on both
> tests.
> I will debug this tonight.
>
> Jörn
--
| Rupert Westenthaler [email protected]
| Bodenlehenstraße 11 ++43-699-11108907
| A-5500 Bischofshofen
