Why is Part-Of-Speech tagging difficult and how to overcome this?



The problem of part-of-speech tagging seems fairly straight-forward.

For a given sentence, identify the part of speech of the various sentences.
Sentence: She loves small animals.
Tags: She(pronoun) loves(noun) small(adjective) animals(noun).

In some other cases, this may be difficult.
Sentence: This can is difficult to open.
While can is usually (auxiliary) verb but here it refers to can( a container) and is thus a noun.

How can we achieve this in part of speech tagging?


Sometimes there are some erroneous results which pop out as a result of POS tagging,but there are a plethora of parsers and taggers you can try to achieve it.

1.You can try the POS tagger in NLTK library of Python
2.NLP & OPENNLP of R (though not that efficient)
3.Stanford parser
4.If you want to find relationships between words,you can try out Dependency Parsing

Above all if you want to identify relationships between words,you can use Dependency Parser (Stanford Parser performs quite good)

Hope this Helps !