Recommended books for Natural Language Processing course

According to the online course of Stanford University, they suggest the following books to be used in NLP class:

• SPEECH and LANGUAGE PROCESSING
• Introduction to Information Retrieval
• Foundations of Statistical Natural Language Processing
• Natural Language Processing with Python

It's now live, you can learn for free, the Standford Online course of Natural Language Processing!!!

Lexalytics: Text Analytics Software

     Lexalytics, Inc. builds text analytics software, and also libraries as well to deal with the natural text from human. It provides the software that can be used to take competitive advantages in businesses such as social media monitoring, marketing intelligence, voice of customers, and enterprise searching. The most interesting thing for me is Tripadvisor is a one of Lexalytics partner, which this website is the world largest travel site, enabling travelers plan and have a perfect trip. Tripadvisor offers trusted advice from real travelers to make beginner travelers plan a suitable solution for themselves.
     In my B.Sc. project, I built a software that have tried to extract sentiment of travelers in positive and negative based on various traveling aspects such as appearance, atmosphere, cleanliness, value, and etc. The hardest things for me are entity mapping, set of vocabularies with feeling for each concerned aspect, error of data, and topic separation. Many things of Lexalytics made me feel curious , and I have tried to study about NLP more at http://www.lexalytics.com/sentiment-extraction-measuring-emotional-tone-text.
     Later, I will summarize important information and write down on this blog.

How to Lemmatization with Wordnet in NLTK

We can convert a word back to its basic form by using wordnet lemmatizer. Moreover, we can specific part of speech to check its form: n->noun, v->verb, a->adjective, r->adverb (if not specify, it will be noun).

from nltk.stem.wordnet import WordNetLemmatizer
l = WordNetLemmatizer()
l.lemmatize('cars')            # car
l.lemmatize('women')           # woman
l.lemmatize('fantasized','v')  # fantasize

So, a word will be checked with wordnet to see its basic form according to part of speech.

Wordnet Interface in Python with NLTK

Currently, my project involves with natural language processing and also wordnet for seeing the relationship between word meanings. In python, wordnet can be used with NLTK. NLTK has a wordnet of Princeton University.The main features that are supported in NLTK as following: synset, lemma, hypernym, hyponym, antonym and path similarity.

Basic use examples:

import nltk
from nltk import wordnet as wn
#looking word's synsets
wn.synsets('beautiful')
#synsets
car = wn.synset('car.n.01')
car.lemmas
car.hypernyms()
car.hyponyms()
#antonyms
car.lemmas[0].antonyms()

See more details at:
http://nltk.googlecode.com/svn/trunk/doc/howto/wordnet.html
http://wordnet.princeton.edu/