Атты І халықаралық конференция ЕҢбектері
жүктеу/скачать 8,57 Mb. Pdf көрінісі
|
| 6.2 Feature extraction
6.2.1 Feature defined According to own characteristics of a Kazakh basic verb , this feature space is defined as: (1) the word, including the current word, the right and left sides of a word. (2) part of speech, including the current word speech, about the two parts of speech information. (3) Affix ingredients, including the current word and the word about the additional ingredient information. (4) Phrase tag that contains the current word and the words to the right and the left two words Phrase marker. This rule-based approach applied to generate the maximum entropy model training corpus, based on Kazakh Linguistics, the feature space show as table 2. Table 2. Feature templates Feature tag Meaning Feature tag meaning w（-1） previous one word POS（+1 ） POS of next one word w（0） the current word POS（+2 ） POS of next two word w（+1） next one word affix（- 1） affix of previous word pos（- 2） POS of previous two word affix（0 ） affix of current word pos（- 1） POS of previous one word affix（1 ） affix of next one word pos（0） POS of the current word жүктеу/скачать 8,57 Mb. Достарыңызбен бөлісу:
|