Morphological disambiguation
by using machine learning
methods
Josef Bušta
19. 11. 2012
19. 11. 2012 Morphological disambiguation by using machine learning methods 2
Input examples – before
Baseline: 93,5197
Total: 4 511 196
k3c4: 4 218 857
k7c7: 292 339
 Czes (465 102 710)
 Context: word1 word2 word3 class word4 word5 word6;
class 2 {k3c4, k7c7}
19. 11. 2012 Morphological disambiguation by using machine learning methods 3
Input examples – now
Baseline: 77,99
Total: 968 228
k3c4: 755 164
K7c7: 213 064 (72.9%)
 s, z, š, ž
 consonant + s
 3 consonants
19. 11. 2012 Morphological disambiguation by using machine learning methods 4
Input examples – word4
Before: Significant
values of attribute:​​
k1c7, k2c7,k3c7
Now: Significant
values of attribute:​​
k1c7, k2c7, k3c7
19. 11. 2012 Morphological disambiguation by using machine learning methods 5
Input examples – word5
Now: Significant
values of attribute:​​
k1c7
Before: Significant
values of attribute:​​
k1c7
19. 11. 2012 Morphological disambiguation by using machine learning methods 6
Results
ZeroR J48 Id3 RF NBTree NB
1000 78 95.5 92.1 95.4 95.7 92.4
PART J48 NB RF NBTree
426 95.6 95.7 93.7 95.4 95.7
The same number of examples in
each class: