← Back to Home

Stock Price Analysis and Prediction

View Code on GitHub

---

Overview

This project presents a comprehensive approach to Stock Price Analysis and Prediction, encompassing:

1. Data Retrieval:
   • Fetching historical stock data using the Financial Modeling Prep (FMP) API.
2. Technical Indicator Computation:
   • Calculating various technical indicators to analyze stock trends and volatility.
3. Machine Learning Models:
   • Implementing models such as Linear Regression, Random Forest, and Long Short-Term Memory (LSTM) networks for stock price prediction.

---

Features

Data Retrieval:

• Financial Modeling Prep (FMP) API Integration:
  • Utilizes FMP’s free API to obtain real-time and historical stock data, supporting over 25,000 stocks across multiple exchanges.

Technical Indicator Computation:

• Moving Averages:
  • Calculates Simple Moving Averages (SMA) and Exponential Moving Averages (EMA) over specified windows to identify price trends.
• Relative Strength Index (RSI):
  • Computes RSI to assess the magnitude of recent price changes and evaluate overbought or oversold conditions.
• Moving Average Convergence Divergence (MACD):
  • Measures the relationship between two EMAs to identify bullish or bearish momentum.
• Bollinger Bands:
  • Calculates Bollinger Bands to analyze price volatility and potential overbought or oversold signals.
• Volatility Metrics:
  • Determines historical volatility and Average True Range (ATR) to gauge market fluctuations.
• On-Balance Volume (OBV):
  • Computes OBV to relate price changes to trading volume, aiding in trend confirmation.

Machine Learning Models:

• Linear Regression:
  • Implements linear regression to model the relationship between lagged closing prices and predict future prices.
• Random Forest Regressor:
  • Utilizes ensemble learning to enhance prediction accuracy and manage overfitting.
• Long Short-Term Memory (LSTM) Networks:
  • Applies LSTM networks to capture temporal dependencies in sequential stock price data for improved forecasting.

---

How It Works

1. Data Retrieval:
   • The stock_history function fetches historical stock data for a specified ticker, year, and month using the FMP API.
2. Technical Indicator Computation:
   • Functions like calculate_moving_averages, calculate_rsi, and calculate_macd compute respective technical indicators using the ta library.
3. Machine Learning Modeling:
   • The prepare_data function structures the data with lag features for regression models.
   • The train_predict_lr_rf function trains and evaluates Linear Regression and Random Forest models.
   • The prepare_lstm_data and train_predict_lstm functions handle data preparation and training for LSTM networks.
4. Prediction and Evaluation:
   • Models are evaluated using metrics such as Root Mean Squared Error (RMSE) to assess prediction accuracy.

---

Technologies

• requests
• BeautifulSoup
• ta (Technical Analysis Library)
• scikit-learn
• TensorFlow

---

← Back to Home