Entity types (PERSON, ORG, LOC, DATE) in NLP - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Entity Types in Named Entity Recognition

Which of the following best describes the entity type ORG in Named Entity Recognition (NER)?

A A person’s name such as 'Alice' or 'Dr. Smith'.

B A location such as 'Paris' or 'Mount Everest'.

C A date or time expression such as 'January 1, 2020'.

D An organization such as 'Google' or 'United Nations'.

💡 Hint

Think about what kind of real-world entities organizations represent.

In NER, ORG stands for organizations like companies, institutions, or groups.

❓ Predict Output

intermediate

Output of NER Entity Extraction

Given the sentence: 'Apple was founded by Steve Jobs in California in 1976.' What entities will be recognized with their correct types?

NLP

sentence = 'Apple was founded by Steve Jobs in California in 1976.'
# Assume a perfect NER model output
entities = [
  ('Apple', 'ORG'),
  ('Steve Jobs', 'PERSON'),
  ('California', 'LOC'),
  ('1976', 'DATE')
]
print(entities)

A [('Apple', 'PERSON'), ('Steve Jobs', 'ORG'), ('California', 'LOC'), ('1976', 'DATE')]

B [('Apple', 'LOC'), ('Steve Jobs', 'PERSON'), ('California', 'ORG'), ('1976', 'DATE')]

C [('Apple', 'ORG'), ('Steve Jobs', 'PERSON'), ('California', 'LOC'), ('1976', 'DATE')]

D [('Apple', 'ORG'), ('Steve Jobs', 'PERSON'), ('California', 'DATE'), ('1976', 'LOC')]

💡 Hint

Think about what each entity represents in real life.

Apple is an organization, Steve Jobs is a person, California is a location, and 1976 is a date.

❓ Model Choice

advanced

Choosing the Best Model for Entity Type Recognition

You want to build a system that identifies PERSON, ORG, LOC, and DATE entities in news articles. Which model type is best suited for this task?

A A linear regression model predicting numeric values.

B A recurrent neural network (RNN) or transformer-based model trained on labeled text data.

C A clustering algorithm like K-means on raw text.

D A convolutional neural network (CNN) trained on image data.

💡 Hint

Think about models designed for sequence and language understanding.

RNNs and transformers are designed to process sequences of text and are commonly used for NER tasks.

❓ Metrics

advanced

Evaluating NER Model Performance

You have a NER model that predicts entity types PERSON, ORG, LOC, and DATE. Which metric best measures how well the model correctly identifies these entities?

A Precision, Recall, and F1-score for each entity type

B Accuracy of predicting the next word

C Mean Squared Error (MSE)

D R-squared value

💡 Hint

Think about metrics used for classification tasks with multiple classes.

Precision, Recall, and F1-score are standard metrics to evaluate classification tasks like NER.

🔧 Debug

expert

Debugging Incorrect Entity Type Predictions

A NER model often labels 'Amazon' as a LOCATION instead of an ORGANIZATION. What is the most likely cause?

A The training data lacks enough examples of 'Amazon' as an organization.

B The model architecture is a CNN instead of an RNN.

C The input text is too short for the model to understand context.

D The model uses precision instead of recall as a metric.

💡 Hint

Think about how the model learns entity meanings from examples.

If the training data does not have enough examples of 'Amazon' as an organization, the model may misclassify it as a location (like the Amazon river).