Emotion Classification Model

This project is a machine learning model that classifies sentences based on the emotion expressed — either positive or negative. It was developed as part of the Technion AI Learning Program to demonstrate a practical end-to-end machine learning pipeline using real-world text data.

Overview

• The dataset includes:
  • Sentences labeled with positive or negative sentiment
  • Country and geographical metadata about the author of each sentence
• The model uses TF-IDF (Term Frequency–Inverse Document Frequency) to convert sentences into numerical vectors
• A K-Nearest Neighbors (KNN) classifier is trained on these vectors to recognize emotional patterns and classify new sentences
• The notebook also includes a simple linear regression demonstration predicting country population based on land area (included for general ML practice)

Machine Learning Pipeline

1. Data Preprocessing

   • Cleaned sentence and country data
   • Transformed text into numeric features using TF-IDF
   • Prepared structured data for modeling

2. Model Training and Evaluation

   • Trained a KNN classifier for binary sentiment classification
   • Evaluated model accuracy on labeled data

3. Bonus: Linear Regression

   • Built a simple linear regression model to predict country population using land size
   • Used as an educational extension unrelated to sentiment classification

Visualizations

• Word clouds showing the most frequent words in positive and negative sentences
• Bar plots for country frequency in the dataset
• Scatter plot showing population vs. land size for regression

Note: There are no visualizations connecting sentiment to country in the current version.

Technologies Used

• Google Colab (Jupyter Notebook)
• Python libraries:
  • pandas, numpy
  • scikit-learn
  • matplotlib, seaborn, wordcloud

How to Run

1. Open the notebook in Google Colab
2. Run all cells in order to:
   • Load and inspect the dataset
   • Train and evaluate the sentiment classifier
   • Generate visualizations
   • Run the regression example

Acknowledgments

This project was completed as part of the Technion AI Learning Program and is intended for educational use to demonstrate core machine learning techniques using text data.