Text classification pipeline in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Model Choice

intermediate

Choosing the best model for short text classification

You have a dataset of short customer reviews labeled as positive or negative. Which model is most suitable to start with for this text classification task?

A A simple Logistic Regression model with TF-IDF features

B A deep Convolutional Neural Network with raw text input

C A K-Means clustering model to group reviews

D A Principal Component Analysis (PCA) model to reduce text dimensions

💡 Hint

Think about models that work well with labeled data and simple features for text.

Logistic Regression with TF-IDF is a strong baseline for text classification, especially for short texts. CNNs require more data and preprocessing. K-Means is unsupervised, and PCA is for dimensionality reduction, not classification.

❓ Predict Output

intermediate

Output of text preprocessing code

What is the output of the following Python code that preprocesses text for classification?

ML Python

import re
text = "Hello World! This is AI-2024."
processed = re.sub(r'[^a-zA-Z ]', '', text).lower().split()
print(processed)

A ['hello', 'world', 'this', 'is', 'ai']

B ['Hello', 'World', 'This', 'is', 'AI']

C ['hello', 'world', 'this', 'is', 'ai2024']

D ['Hello', 'World', 'This', 'is', 'AI2024']

💡 Hint

Look at how non-letter characters are removed and text is lowered.

The regex removes all characters except letters and spaces, so 'AI-2024' becomes 'AI'. Then the text is converted to lowercase and split into words.

❓ Hyperparameter

advanced

Selecting the best hyperparameter for TF-IDF vectorizer

You want to improve your text classification model by tuning the TF-IDF vectorizer. Which hyperparameter controls the maximum number of features (words) to keep?

A min_df

B max_features

C ngram_range

D stop_words

💡 Hint

This parameter limits the vocabulary size.

max_features sets the maximum number of words to keep based on term frequency. min_df controls minimum document frequency, ngram_range controls word group sizes, and stop_words removes common words.

❓ Metrics

advanced

Interpreting classification report metrics

Your text classification model outputs the following metrics: precision=0.8, recall=0.5, accuracy=0.75. What does the low recall indicate?

A The model has perfect predictions

B The model predicts too many false positives

C The model has balanced errors between classes

D The model misses many positive examples (false negatives are high)

💡 Hint

Recall measures how many actual positives are found.

Low recall means the model fails to identify many true positive cases, resulting in many false negatives. Precision relates to false positives, accuracy is overall correctness.

🔧 Debug

expert

Debugging a text classification pipeline error

You run this code snippet for training a text classification model but get a ValueError: Found input variables with inconsistent numbers of samples. What is the cause?

ML Python

from sklearn.feature_extraction.text import CountVectorizer
from sklearn.linear_model import LogisticRegression
texts = ['good', 'bad', 'average']
labels = [1, 0]
vectorizer = CountVectorizer()
X = vectorizer.fit_transform(texts)
model = LogisticRegression()
model.fit(X, labels)

A LogisticRegression requires labels to be strings

B CountVectorizer cannot process single words

C The labels list length does not match the number of text samples

D fit_transform returns a dense matrix, but model expects sparse

💡 Hint

Check if the number of labels matches the number of texts.

The error occurs because texts has 3 samples but labels has only 2, causing mismatch in training data and labels.