Text Classification with Logistic Regression for Sentiment Analysis

Project Overview

This project involves analyzing customer reviews to classify them as positive or negative using Logistic Regression. The workflow includes text preprocessing, feature extraction, training a model, making predictions, and evaluating its performance.

Dataset

The dataset consists of customer reviews labeled with sentiment scores:

• Review: The text of the customer’s review
• Sentiment: The target variable (1 = Positive, 0 = Negative)

Dataset file: customer_reviews_sentiment.csv

Requirements

Install the necessary dependencies before running the code:

pip install pandas numpy scikit-learn nltk

Workflow

1. Data Preprocessing

• Load the dataset
• Handle missing values
• Convert text to lowercase
• Remove special characters and punctuation
• Remove stop words (using NLTK or SpaCy)
• Tokenize the text

2. Feature Extraction

• Convert text into numerical features using:
  • Bag of Words (BoW) or
  • TF-IDF (Term Frequency-Inverse Document Frequency)
• Split the dataset into training (80%) and testing (20%) sets

3. Train a Logistic Regression Model

• Train a Logistic Regression classifier on extracted features
• Tune hyperparameters (experiment with regularization parameter C)

4. Make Predictions

• Predict the sentiment for the following reviews:
  • "This product is amazing! I love it."
  • "It broke after one use, completely disappointed."

5. Model Evaluation

• Compute accuracy on the test dataset
• Generate confusion matrix and classification report (precision, recall, F1-score)

6. Model Improvements

• Experiment with other classifiers (e.g., Naive Bayes, SVM)
• Compare their performance with Logistic Regression

Deliverables

1. Preprocessed Dataset: Cleaned text data
2. Feature-Engineered Dataset: Extracted numerical features
3. Trained Model: Logistic Regression with optimized hyperparameters
4. Model Evaluation: Accuracy, confusion matrix, classification report
5. Sample Predictions: Results for provided test cases
6. Model Comparison: Performance of alternative classifiers

How to Run the Code

1. Ensure the dataset is available as customer_reviews_sentiment.csv
2. Run the preprocessing and feature extraction scripts
3. Train the Logistic Regression model
4. Evaluate performance and compare models

Author

Pedahel Emmanuel Kojo
Senior Software Engineer, Machine Learning Engineer at CSP