Emotion Classification

This project focuses on emotion classification using machine learning and deep learning techniques. It preprocesses image data, applies data augmentation, and uses various classifiers, including tree-based models and deep learning approaches.

Features

• Data Preprocessing: Image transformations, standardization, and balancing techniques like SMOTE.
• Feature Extraction: PCA for dimensionality reduction.
• Machine Learning Models: ExtraTrees, XGBoost, and Optuna hyperparameter tuning.
• Deep Learning Models: Neural networks (if applicable in the notebook).

Requirements

Install dependencies using:

pip install -r requirements.txt

Usage

Run the Jupyter Notebook for data preprocessing, model training, and evaluation.

File Structure

• emotion_classification.ipynb: Main notebook for data processing and training.
• requirements.txt: Required libraries. <<<<<<< HEAD

=======

Model Performance

Below are the assumed performance metrics of the three models on the test set:

Model	Accuracy	F1-Score (Average)	Precision (Average)	Recall (Average)
XGBoost	81%	0.8	0.82	0.8
ExtraTrees	73%	0.7	0.70	0.73

f1d74716d616e8ad67c295ce52a6a013eabf9297