#  Competition Description

Human emotions are complex and often expressed through informal text, slang, and emojis—especially on social media platforms.

In this competition, you are provided with a dataset of tweets labeled with one of six emotions:

Your task is to build a classification model that predicts the correct emotion for each tweet in the test dataset.

---

##  Dataset Files

You are provided with:

- `train.csv` – Training data with emotion labels
- `test.csv` – Test data without labels
- `sample_submission.csv` – Example submission format

---

##  Challenge Details

This dataset includes:
- Informal language
- Emoji usage
- Mixed emotional expressions
- Real-world noisy text

Participants may use:
- Traditional ML methods (TF-IDF, Logistic Regression, SVM)
- Deep Learning (LSTM, CNN)
- Transformer-based models (BERT, DistilBERT, etc.)

External publicly available pretrained models are allowed.

---

##  Objective

Your objective is to:

1. Preprocess text effectively  
2. Engineer meaningful features  
3. Train a robust multi-class classifier  
4. Maximize prediction accuracy on the hidden test set  

Final rankings will be based on performance on the **Private Leaderboard**.

---

We encourage:
- Clean code
- Reproducibility
- Creative feature engineering
- Ethical AI practices

Let the competition begin!