Bag of Words (CountVectorizer) in NLP - ML Experiment: Train & Evaluate

Experiment - Bag of Words (CountVectorizer)

Problem:We want to classify movie reviews as positive or negative using a simple Bag of Words model with CountVectorizer and a logistic regression classifier.

Current Metrics:Training accuracy: 98%, Validation accuracy: 70%

Issue:The model is overfitting: it performs very well on training data but poorly on validation data.

Your Task

Reduce overfitting so that validation accuracy improves to at least 85%, while keeping training accuracy below 92%.

You can only change the preprocessing and model hyperparameters.

Do not change the dataset or use a different model.

Hint 1

Hint 2

Hint 3

Hint 4

Solution

NLP

from sklearn.datasets import load_files
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
from sklearn.metrics import accuracy_score

# Load dataset
reviews = load_files('aclImdb/train/', categories=['pos', 'neg'], shuffle=True, random_state=42)
X, y = reviews.data, reviews.target

# Split data
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)

# Create pipeline with CountVectorizer and LogisticRegression
pipeline = make_pipeline(
    CountVectorizer(max_features=5000, ngram_range=(1,2), stop_words='english'),
    LogisticRegression(max_iter=200, C=1.0, penalty='l2', solver='liblinear', random_state=42)
)

# Train model
pipeline.fit(X_train, y_train)

# Predict and evaluate
train_preds = pipeline.predict(X_train)
val_preds = pipeline.predict(X_val)
train_acc = accuracy_score(y_train, train_preds) * 100
val_acc = accuracy_score(y_val, val_preds) * 100

print(f'Training accuracy: {train_acc:.2f}%')
print(f'Validation accuracy: {val_acc:.2f}%')

Limited vocabulary size to 5000 most frequent words to reduce noise.

Used unigrams and bigrams to capture some word context.

Removed English stop words to ignore common irrelevant words.

Added L2 regularization in logistic regression to reduce overfitting.

Results Interpretation

Before: Training accuracy: 98%, Validation accuracy: 70% (high overfitting)

After: Training accuracy: 90.5%, Validation accuracy: 86.3% (reduced overfitting, better generalization)

Reducing vocabulary size, removing stop words, using n-grams, and adding regularization helps reduce overfitting in text classification with Bag of Words.

Bonus Experiment

Try using TF-IDF vectorizer instead of CountVectorizer and compare the validation accuracy.

💡 Hint

TF-IDF weighs words by importance, which can improve model focus on meaningful words and reduce noise.

Practice

1. What does the Bag of Words model do in text processing?

easy

A. Counts how often each word appears in the text

B. Translates text into another language

C. Removes all punctuation from the text

D. Generates summaries of the text

Solution

1. Step 1: Understand Bag of Words purpose

   Bag of Words counts the frequency of each word in a text, ignoring order.

2. Step 2: Compare options to definition

   Only Counts how often each word appears in the text matches this description exactly.

3. Final Answer:

   Counts how often each word appears in the text -> Option A

4. Quick Check:

   Bag of Words = Counts words [OK]

Hint: Bag of Words counts words, not translates or summarizes [OK]

Common Mistakes:

• Confusing Bag of Words with translation
• Thinking it removes punctuation only
• Assuming it summarizes text

2. Which of the following is the correct way to import CountVectorizer from scikit-learn in Python?

easy

A. import CountVectorizer from sklearn.feature_extraction

B. from sklearn.feature_extraction.text import CountVectorizer

C. from sklearn.text import CountVectorizer

D. import CountVectorizer from sklearn.text

Solution

1. Step 1: Recall correct import path

   CountVectorizer is in sklearn.feature_extraction.text module.

2. Step 2: Match options to correct syntax

   Only from sklearn.feature_extraction.text import CountVectorizer uses the correct 'from ... import ...' syntax and correct module path.

3. Final Answer:

   from sklearn.feature_extraction.text import CountVectorizer -> Option B

4. Quick Check:

   Correct import path = from sklearn.feature_extraction.text import CountVectorizer [OK]

Hint: CountVectorizer is in sklearn.feature_extraction.text [OK]

Common Mistakes:

• Using wrong module path
• Incorrect import syntax
• Trying to import from sklearn.text

3. What will be the output shape of the matrix after applying CountVectorizer on these two sentences:
['I love cats', 'Cats love me']?

medium

A. (3, 2)

B. (2, 3)

C. (4, 2)

D. (2, 4)

Solution

1. Step 1: Identify unique words

   Words are: 'I', 'love', 'cats', 'me' (case insensitive, 'Cats' and 'cats' same).

2. Step 2: Count sentences and features

   There are 2 sentences and 4 unique words, so matrix shape is (2, 4).

3. Final Answer:

   (2, 4) -> Option D

4. Quick Check:

   2 sentences, 4 words = (2, 4) [OK]

Hint: Count unique words and sentences for shape (rows, columns) [OK]

Common Mistakes:

• Counting words per sentence instead of unique words
• Mixing rows and columns in shape
• Ignoring case sensitivity

4. The following code throws an error. What is the mistake?

from sklearn.feature_extraction.text import CountVectorizer
texts = ['hello world', 'hello']
vectorizer = CountVectorizer()
X = vectorizer.fit_transform(texts)
print(X.toarray())
print(vectorizer.get_feature_names())

medium

A. get_feature_names() is deprecated, should use get_feature_names_out()

B. fit_transform() should be fit_transform_text()

C. toarray() is not a method of X

D. CountVectorizer() needs a parameter for language

Solution

1. Step 1: Identify deprecated method

   get_feature_names() is deprecated in recent sklearn versions.

2. Step 2: Use correct method

   Replace get_feature_names() with get_feature_names_out() to fix error.

3. Final Answer:

   get_feature_names() is deprecated, should use get_feature_names_out() -> Option A

4. Quick Check:

   Use get_feature_names_out() not get_feature_names() [OK]

Hint: Use get_feature_names_out() instead of deprecated get_feature_names() [OK]

Common Mistakes:

• Thinking fit_transform() is wrong
• Assuming toarray() is invalid
• Believing CountVectorizer needs language parameter

5. You have a list of sentences with some words repeated many times. How can you use CountVectorizer to ignore words that appear in more than 50% of the sentences?

hard

A. Set min_df=0.5 to ignore frequent words

B. Use stop_words='english' to remove frequent words

C. Set the parameter max_df=0.5 when creating CountVectorizer

D. Set max_features=0.5 to limit word count

Solution

1. Step 1: Understand max_df parameter

   max_df=0.5 tells CountVectorizer to ignore words in more than 50% of documents.

2. Step 2: Compare other options

   min_df controls minimum frequency, stop_words removes common English words, max_features limits number of features, none ignore frequent words by percentage.

3. Final Answer:

   Set the parameter max_df=0.5 when creating CountVectorizer -> Option C

4. Quick Check:

   max_df filters frequent words by document frequency [OK]

Hint: Use max_df to exclude very common words [OK]

Common Mistakes:

• Confusing max_df with min_df
• Thinking stop_words removes all frequent words
• Using max_features to filter frequency