Emotion Detection using Deep Neural Networks

This project implements an end-to-end Emotion Detection system using Deep Neural Networks in PyTorch. The model classifies human emotions from text by learning contextual patterns using an LSTM-based architecture.

Problem Statement

Understanding human emotions from text is an important task in Natural Language Processing (NLP).This project aims to build a Deep Learning model that can automatically classify emotions from textual data,which can be useful in applications such as chatbots, customer feedback analysis, and mental health monitoring.

Dataset Description

The project uses a real-world multi-class emotion dataset consisting of text samples labeled with emotion classes. The dataset is split into:

• Training set (training.csv)
• Validation set (validation.csv)
• Test set (test.csv)

Each file contains:

• text: Input sentence
• label: Emotion class (0–5)

Model Architecture

The Emotion Detection model is based on a Deep Neural Network with the following architecture:

Input Text

• Tokenization & Padding
• Embedding Layer
• LSTM Layer
• Fully Connected Layer
• Emotion Class Prediction

Technologies & Tools Used

• Python
• PyTorch
• Pandas
• NumPy
• Deep Learning
• Natural Language Processing (NLP)

Data Preprocessing Steps

The following preprocessing steps were applied to the text data:

1. Text normalization (lowercasing and punctuation removal)
2. Tokenization
3. Vocabulary creation using training data only
4. Conversion of text into numerical sequences
5. Padding sequences to a fixed length

Training Details

• Model : LSTM-based Neural Network
• Loss Function: CrossEntropyLoss
• Optimizer: Adam
• Batch Size: 32
• Sequence Length: 50
• Number of Epochs: 15

Evaluation Results

The model was evaluated on a validation dataset to measure its generalization performance.

• Validation Accuracy: ~85% The results indicate that the model learns meaningful emotional patterns from text data.

How to Run the Project

1. Clone the repository: git clone https://github.com/Tirth9978/emotion-detection-dnn.git
2. Install dependencies: pip install -r requirements.txt
3. Run the Main.ipynb File

Sample Predictions

1. Input: "I feel lonely and exhausted" Predicted Emotion: Sadness
2. Input: "Today is the best day of my life" Predicted Emotion: Joy

---

Project Structure

emotion-detection-dnn/
│
├── data/
├── src/
├── notebooks/
├── requirements.txt
├── model.pt
├── README.md

Future Improvements

• Add the attention mechanish to improve the performance
• Use Bidirectional LSTM
• Integrate Transformer-based models
• Deploy the model as a web application

License

This project is open-source and available under the MIT License.

👤 Author

Tirth Patel
Aspiring Deep Learning Engineer | NLP & AI Enthusiast

I am a passionate and disciplined learner with a strong interest in Deep Learning, Natural Language Processing (NLP), and Artificial Intelligence. I actively work on real-world machine learning projects to strengthen my practical understanding of neural networks and modern AI systems.

This project reflects my hands-on experience with:

• Building end-to-end Deep Learning pipelines
• Working with real-world datasets
• Designing and training LSTM-based neural networks
• Implementing data preprocessing, model training, validation, and inference using PyTorch

I believe in learning concepts deeply by implementing them from scratch rather than relying only on high-level abstractions. My long-term goal is to contribute to impactful AI systems and research-driven engineering projects while continuously improving my problem-solving and software development skills.

📌 GitHub Profile:
🔗 https://github.com/Tirth9978

📌 Areas of Interest:

• Deep Learning
• Natural Language Processing (NLP)
• Neural Networks
• PyTorch
• Machine Learning
• AI Research

📌 Project Motivation:
This project was built as a resume-ready, real-world Deep Learning application to demonstrate my understanding of neural networks, text representation, and model evaluation in a structured and professional manner.