Stock Price Prediction Model: An Overview of [DP Fincial]

This publication documents a Deep Learning model developed to predict future stock price trends using historical data.

This model is designed to help investors make informed and data-driven financial decisions.

1. Objective of the Model

The main objective of this project is to predict the closing price of major stocks or indices (like Nifty 50) listed in the Indian stock market for the next 7 days through time-series analysis .

2. Dataset & Source

• Data Source: Publicly available data from major Indian stock exchanges (like NSE).
• Period: Daily stock data for last 5 years (2020-2025).
• Features used: Open, High, Low, Close prices and trading volume.
• Data Pre-processing: The data is normalized using MinMaxScaler before model training.

3. Model and Methodology

This model is based on the LSTM (Long Short-Term Memory) neural network , which is highly effective in recognizing patterns and long-term dependencies in time-series data.

• Algorithm: LSTM (Deep Learning)
• Architecture: [2 LSTM Layers] + [1 Dense Output Layer]
• Training: The model was trained for [50 Epochs], with 80% of the data used for training and 20% for testing.
• Loss Function: Mean Squared Error (MSE)

4. Results and Evaluation

During testing, the model showed the following key metrics:

• RMSE (Root Mean Squared Error): It shows how far the predicted value is, on average, from the actual value.
• MAPE (Mean Absolute Percentage Error): [Example: 3% to 5%]. It shows the prediction error in percentage.

The model's results show that it is able to track stock price trends with low error, proving its practical utility.

5. Conclusion and Future Work

This LSTM-based model provides a strong foundation for stock prediction. Future work will include the following steps:

1. Incorporating market sentiment analysis from news headlines and social media data.
2. Experimenting with other advanced model architectures like Transformer or Conv1D for better performance.