Portfolio Optimization Using Market Correlations

Forecasting stock prices using ARIMA, LSTM, and Prophet time-series models.

This research project provides a meticulous assessment of the forecasting prowess of three renowned time series models using ARIMA, LSTM, and Prophet applied to stock prices of major companies (Apple, Microsoft, NVIDIA) from prominent indices such as NYSE, NASDAQ, and S&P 500.

Optimizing Investment Portfolios using Time Series Forecasting

A comparative analysis of ARIMA, LSTM, and Prophet models for stock price prediction across major market indices including NASDAQ, NYSE, and S&P 500.

Overview

This project evaluates the forecasting performance of three popular time series models:

  • ARIMA
  • LSTM
  • Prophet

The study also includes sector-wise stock analysis and visualization of prediction trends for technology companies such as Apple, Microsoft, and Nvidia.

Key Features

  • Time series forecasting on historical stock market data
  • Comparative analysis using RMSE and MSE
  • Sectoral trend analysis
  • Interactive visualizations using Plotly
  • Model performance benchmarking

Results

  • LSTM achieved the highest prediction accuracy with the lowest RMSE and MSE.
  • Prophet performed moderately well for seasonal trend analysis.
  • ARIMA showed limitations with highly volatile and non-linear stock movements.

Tech Stack

  • Python
  • Pandas
  • NumPy
  • Scikit-learn
  • TensorFlow / Keras
  • Statsmodels
  • Facebook Prophet
  • Plotly
  • Matplotlib

Dataset

Historical stock market dataset collected from Kaggle containing stocks listed on:

  • NASDAQ
  • NYSE
  • S&P 500