Sentiment Analysis Project

This project builds models to classify consumer opinions into negative, neutral, or positive sentiment. The following models are implemented:

• DistilBERT (fine-tuned transformer)
• Logistic Regression (TF-IDF features)
• SVM (LinearSVC) (TF-IDF features)

This code is part of a final project for the TextMining course, therefore comments and descriptions in the file are in polish language.

Contents

1. Data Preparation

   • Text preprocessing: cleaning, tokenization, stopword removal, lemmatization.
   • Class balancing using oversampling.

2. Model Training

   • Logistic Regression and SVM trained on TF-IDF vectors.
   • DistilBERT fine-tuned on the processed data.

3. Evaluation

   • F1-score, precision, recall, and accuracy are calculated.
   • Confusion matrices and comparisons of model performance are included.