Named Entity Recognition (NER) is a vital task in Natural Language Processing (NLP), particularly in applications that involve information extraction, question answering, or text summarization. It involves identifying and classifying named entities in text, such as people, organizations, locations, dates, and more.
In this article, we will explore how to identify and classify named entities using Python. We will focus on using the popular NLP library called Natural Language Toolkit (NLTK) along with the Stanford Named Entity Recognizer (NER) to perform this task effectively.
Before starting, make sure you have Python installed on your system. To install the NLTK library, you can use the following command:
pip install nltk
Additionally, we will use the Stanford Named Entity Recognizer (NER), which requires Java. You can download it by visiting the Stanford NLP website and following the installation instructions.
Once you have installed NLTK and downloaded the Stanford NER, you need to import the necessary libraries and load the pre-trained NER model. Here's how you can do it:
import nltk
from nltk.tag import StanfordNERTagger
# Set the path to your Java JDK and Stanford NER models
java_path = "/path_to_java/bin/java"
stanford_path = "/path_to_stanford_ner"
# Set the classpath to the Stanford NER JAR file
stanford_jar = stanford_path + "/stanford-ner.jar"
# Set the path to the pre-trained model
model = stanford_path + "/classifiers/english.all.3class.distsim.crf.ser.gz"
# Initializing the Stanford NER Tagger
ner_tagger = StanfordNERTagger(model, stanford_jar, java_path=java_path)
To identify and classify named entities, we need some text to work with. Let's assume we have a text document called "sample.txt". To load and process the text using the Stanford NER tagger, you can use the following code:
# Load the text document
with open("sample.txt", "r") as file:
text = file.read()
# Tokenize the text into sentences
sentences = nltk.sent_tokenize(text)
# Perform NER on each sentence
for sentence in sentences:
words = nltk.word_tokenize(sentence)
tags = ner_tagger.tag(words)
print(tags)
The code above loads the text file, tokenizes it into sentences using NLTK, and then performs NER on each sentence using the Stanford NER tagger. The tags
variable contains the named entity labels assigned to each word/token in the sentence.
To extract and classify the named entities from the text, we can iterate through the tagged words and filter out the relevant entities. Here's an example of how to do it:
# Perform NER and extract named entities
named_entities = []
for sentence in sentences:
words = nltk.word_tokenize(sentence)
tags = ner_tagger.tag(words)
named_entities.extend([(word, tag) for word, tag in tags if tag != "O"])
# Classify named entities into predefined categories
person_entities = [word for word, tag in named_entities if tag == "PERSON"]
organization_entities = [word for word, tag in named_entities if tag == "ORGANIZATION"]
location_entities = [word for word, tag in named_entities if tag == "LOCATION"]
In the code above, the named entities are filtered based on their tag. In this example, we extract three types of named entities: person, organization, and location. You can customize this filtering based on your specific use case.
Named Entity Recognition plays a crucial role in various NLP applications. In this article, we have explored how to identify and classify named entities using Python and the Stanford NER tagger. By understanding and utilizing NER, you can extract valuable information from text documents, enabling more advanced analysis and decision-making in your NLP projects.
noob to master © copyleft