How to use POS tags as the features

2 次查看（过去 30 天）

Saugata Bose 2019-9-11

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/479821-how-to-use-pos-tags-as-the-features

编辑： Saugata Bose 2019-9-11

I am looking forward to know how could I use POS tags as the features. Adding partOfSpeechDetails after tokenizing the document has tagged every word with its respective POS. But how could I take these tags as the features to fed into a classifier?

As an example, for the sentence, "hello. how are you?", I got the POS details as the following:

Token     DocumentNumber    SentenceNumber    LineNumber       Type        Language     PartOfSpeech 
    _______    ______________    ______________    __________    ___________    ________    ______________
    "hello"          1                 1               1         letters           en       interjection  
    "."              1                 1               1         punctuation       en       punctuation   
    "how"            1                 2               1         letters           en       adverb        
    "are"            1                 2               1         letters           en       auxiliary-verb
    "you"            1                 2               1         letters           en       pronoun       
    "?"              1                 2               1         punctuation       en       punctuation  

Now, how could I take the PartOfSpeech columns as a feature for the sentence? And do u shed some light on how many part of speeches are avilable in Matlab?

I am looking for your advice in this regard.

Thanks,