4 These are the linguistic data for the Apertium Basque--Spanish machine translator.
5 It translates only in the direction Basque--Spanish.
6 You need apertium-2.0 and lttoolbox-2.0 to use this translator.
9 To compile the linguistical data simply do:
13 to generate a Makefile file and then
17 inside of this directory.
22 To use this language-pair package with apertium YOU DO NOT NEED TO
23 RETRAIN THE TAGGER. Probabilities and auxiliary data are provided for
24 both the en-ca and the ca-en translation directions which should be
25 acceptable for most applications, and should work even if you change
26 the dictionaries in a reasonably way.
28 If for some reason you need to retrain the tagger (for example, you
29 have made really extensive changes to the dictionaries such as
30 creating new lexical categories), you have three alternatives:
32 * To perform a supervised training:
34 To this end tagged corpora is provided, but tagged corpora
35 (eu-tagger-data/eu.tagged ) could be
36 obsolete for some words. If this is the case, the tagger training
37 program will show you where the problems are and you will need
38 to solve them by hand. Be sure to solve the problems by modifying
39 ONLY the .tagged file, NEVER the .untagged file that is
40 automatically generated.
42 The supervised training is done by typing:
44 make -f eu-es-supervised.make (for the Basque part-of-speech tagger)
45 * To perform an unsupervised training:
47 For this purpose you will need to assemble a large (hundreds of
48 thousand of words) plain-text corpus for each language (for example,
49 using a robot to harvest text from online newspapers) and put them in
50 the proper place, for instance eu-tagger-data/eu.crp.txt. This type
51 of training does not need human
52 intervention but, as expected, results will be less adequate than
53 those obtained with the supervised training.
55 The unsupervised training is done through the iterative Baum-Welch
56 algorithm. By default the number of iterations is set to 8, but you
57 can change this value by editing the Makefile and changing the
58 value of TAGGER_UNSUPERVISED_ITERATIONS.
60 The unsupervised training is done by typing:
62 make -f eu-es-unsupervised.make (for the Basque part-of-speech tagger)
64 This is the training method followed to train the basque tagger.
67 * To perform an unsupervised training by using target-language
68 information and the rest of the modules of the Apertium MT engine:
70 To do so you need large plain-text corpora on both languages. Please
71 download the apertium-tagger-training-tools package and follow the
72 instructions provided there.