📈 MNC Stock Market Price Prediction

Time-series stock price forecasting using Random Forest, LSTM (Deep Learning), and ARIMA on IBM historical stock data.

📌 Overview

This project applies three distinct forecasting approaches to IBM stock price data, comparing their predictive power on closing prices:

• Random Forest Regressor — ensemble tree-based approach
• LSTM Neural Network — 4-layer stacked LSTM with Dropout regularisation
• ARIMA / Auto-ARIMA — classical time-series decomposition and forecasting

🗂️ Repository Structure

MNC-Stock-Market-Price-Prediction/
├── MNC_Stock_Market_Price_Prediction.ipynb   # Full analysis and modelling notebook
├── mnc_stock_market_price_prediction.py      # Python script version
├── IBM Stock Data.csv                        # Historical IBM stock data (OHLCV)
├── requirements.txt                          # Python dependencies
└── README.md                                 # This file

📊 Dataset

Feature	Description
Date	Trading date
Open	Opening price (USD)
High	Intraday high price (USD)
Low	Intraday low price (USD)
Close	Closing price (USD) — target variable
Volume	Number of shares traded
Name	Ticker symbol (IBM)

🧠 Model Architectures

🌲 Random Forest

• 100 decision trees with bootstrap aggregation
• Features: lag prices, rolling mean, rolling std
• Evaluated on 20% holdout set

🔁 LSTM (Long Short-Term Memory)

Layer 1: LSTM(50, return_sequences=True) + Dropout(0.2)
Layer 2: LSTM(50, return_sequences=True) + Dropout(0.2)
Layer 3: LSTM(50, return_sequences=True) + Dropout(0.2)
Layer 4: LSTM(50) + Dropout(0.2)
Output:  Dense(1)

• Optimizer: Adam | Loss: Mean Squared Error
• Look-back window: 60 days

📉 ARIMA

• Augmented Dickey-Fuller (ADF) test for stationarity
• Seasonal decomposition (trend, seasonal, residual)
• Auto-ARIMA for optimal (p, d, q) selection

📈 Results

Model	RMSE
Random Forest	~4.2
LSTM	~3.1
ARIMA	~5.8

LSTM achieves the lowest RMSE, capturing long-term temporal dependencies in the stock price sequence.

🔮 Future Work

• Add Bidirectional LSTM and GRU variants
• Incorporate sentiment analysis from financial news (NLP)
• Multi-stock portfolio prediction
• Real-time prediction dashboard with Streamlit
• Add technical indicators: RSI, MACD, Bollinger Bands

⚙️ Setup

pip install -r requirements.txt
jupyter notebook MNC_Stock_Market_Price_Prediction.ipynb

👩‍💻 Author

Devyani Deore — github.com/DevyaniD19

📄 License

MIT License — see LICENSE for details.