NER with spaCy in NLP - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Named Entity Recognition with spaCy

What is the output of the following code snippet that uses spaCy to detect named entities in a sentence?

NLP

import spacy
nlp = spacy.load('en_core_web_sm')
doc = nlp('Apple is looking at buying U.K. startup for $1 billion')
entities = [(ent.text, ent.label_) for ent in doc.ents]
print(entities)

A [('Apple', 'ORG'), ('U.K.', 'GPE'), ('$1 billion', 'MONEY')]

B [('Apple', 'PERSON'), ('U.K.', 'LOC'), ('$1 billion', 'MONEY')]

C [('Apple', 'ORG'), ('startup', 'ORG'), ('$1 billion', 'MONEY')]

D [('Apple', 'ORG'), ('U.K.', 'ORG'), ('$1 billion', 'MONEY')]

💡 Hint

Look at common entity labels spaCy uses for organizations, geopolitical entities, and money.

spaCy labels 'Apple' as an organization (ORG), 'U.K.' as a geopolitical entity (GPE), and '$1 billion' as money (MONEY).

❓ Model Choice

intermediate

Choosing the Correct spaCy Model for NER

Which spaCy model is best suited for performing Named Entity Recognition on English text with a balance of speed and accuracy?

A en_core_web_md

B en_vectors_web_lg

C en_core_web_lg

D en_core_web_sm

💡 Hint

Consider models that include trained pipelines for NER and balance size and speed.

en_core_web_md provides a good balance of speed and accuracy with medium-sized vectors and trained NER pipeline.

❓ Hyperparameter

advanced

Effect of Batch Size in spaCy NER Training

During training a spaCy NER model, what is the effect of increasing the batch size parameter?

A It slows down training and always improves model accuracy.

B It causes the model to overfit immediately.

C It has no effect on training speed or accuracy.

D It speeds up training but may reduce model accuracy due to less frequent weight updates.

💡 Hint

Think about how batch size affects gradient updates and training speed.

Larger batch sizes process more examples before updating weights, speeding training but possibly reducing accuracy due to less frequent updates.

❓ Metrics

advanced

Evaluating spaCy NER Model Performance

Which metric is most appropriate to evaluate the quality of a spaCy Named Entity Recognition model on a test dataset?

A Accuracy of token classification

B Mean Squared Error

C F1-score

D Perplexity

💡 Hint

NER is a sequence labeling task with imbalanced classes; consider precision and recall.

F1-score balances precision and recall, making it ideal for evaluating NER models where both false positives and false negatives matter.

🔧 Debug

expert

Debugging spaCy NER Model Training Error

You run spaCy NER training code but get this error: 'ValueError: [E088] The component 'ner' is not initialized. Call 'nlp.begin_training()' first.' Which code snippet fixes this error?

A

nlp = spacy.load('en_core_web_sm')
ner = nlp.get_pipe('ner')
optimizer = nlp.begin_training()

B

nlp = spacy.blank('en')
ner = nlp.add_pipe('ner')
optimizer = nlp.begin_training()

C

nlp = spacy.load('en_core_web_sm')
optimizer = nlp.create_optimizer()

D

nlp = spacy.blank('en')
optimizer = nlp.begin_training()

💡 Hint

The error means the 'ner' component is missing or not initialized properly.

When creating a blank model, you must add the 'ner' pipe before calling begin_training(). Option B does this correctly.