📈 Yahoo-Finance-Stock-Market-Prediction

Time Series Analysis using LSTM

---

📖 Overview

This project predicts stock market prices using Long Short-Term Memory (LSTM) networks.
It utilizes Yahoo Finance stock market data, preprocesses it, and trains an LSTM model to predict future stock prices.

🔹 Key Features:

✔️ Data Preprocessing – Cleans & normalizes stock prices
✔️ LSTM Model – Predicts stock prices based on historical data
✔️ Visualization – Plots training loss & actual vs predicted prices
✔️ W&B Integration – Logs training metrics & visualizations

---

📂 Project Structure

📂 yahoo_stock_lstm/
│── 📄 dataload.py       # Loads & preprocesses the dataset
│── 📄 model.py          # Defines the LSTM model
│── 📄 train.py          # Trains the LSTM model
│── 📄 inference.py      # Runs inference (predictions)
│── 📄 README.md         # Project documentation
│── 📄 requirements.txt  # Dependencies (optional)
│── 📄 training_loss.png # Training loss graph
│── 📄 lstm_prediction.png # Prediction results

---

🛠️ Setup & Installation

1️⃣ Clone the repository

git clone https://github.com/your-username/Yahoo-Finance-Stock-Market-Prediction.git
cd Yahoo-Finance-Stock-Market-Prediction

2️⃣ Install dependencies

pip install -r requirements.txt

3️⃣ Set up Weights & Biases (W&B)
Login to W&B with your API key:

wandb login YOUR_WANDB_API_KEY

---

🧩 Usage

1️⃣ Train the LSTM Model

python train.py

✅ This will:
✔️ Load the dataset
✔️ Train the LSTM model
✔️ Save the trained model as lstm_model.h5
✔️ Log metrics to W&B Training Dashboard

🔗 My Training Logs: View Here

---

2️⃣ Run Inference & Predict Stock Prices

python inference.py

✅ This will:
✔️ Load the trained model
✔️ Make predictions on the test set
✔️ Plot actual vs predicted stock prices

🔗 My Inference Logs: View Here

---

📊 Results

📌 Training Loss Plot:

📌 Predicted vs Actual Stock Prices:

🔗 Full W&B Project Dashboard:
👉 View All Logs & Visualizations

---

🛠️ Technologies Used

• Python 🐍
• TensorFlow/Keras 🔥
• Scikit-Learn 📊
• Matplotlib & Seaborn 📉
• Weights & Biases (W&B) 🚀

---

📌 Future Improvements

✅ Implement W&B Sweeps for hyperparameter tuning
✅ Compare GRU-based models
✅ Enhance feature engineering with technical indicators

---

📜 License

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

---

📬 Contact

👤 Rohan Shenoy
📧 Email: roshenoy30@gmail.com
🔗 GitHub: rohanshenoy30

---