📈 Market Predictor – Forecasting European Stock Indices with Machine Learning

🎯 Project Objective

This project investigates whether it is possible to predict — better than random chance — the future movement of major European stock indices using machine learning models.

We focus on the following three benchmark indices:

• 🇫🇷 CAC 40 (France)
• 🇪🇺 STOXX Europe 600 (Pan-European, includes UK)
• 🇪🇺 EURO STOXX 50 (Eurozone only)

---

🧠 Research Background & Literature

This project builds on recent research in financial machine learning, which highlights the role of technical indicators, macroeconomic variables, and investor sentiment in forecasting market behavior.

🔬 Key References

• Kumbure et al. (2022) – A comprehensive review of 138 ML-based stock forecasting studies (2000–2019). Most models use technical indicators (RSI, SMA, MACD) and techniques such as SVM, neural networks, and LSTM.
• Liu & Long (2020) – A hybrid architecture combining Empirical Wavelet Transform, deep LSTM, and Extreme Learning Machine for predicting daily closing prices of major indices. Outperforms standard LSTM and random forests.
• Lin et al. (2021) – Classify short-term market direction (up/down) using candlestick patterns and 21 technical indicators with various ML algorithms (logistic regression, k-NN, GBDT, LSTM).
• Ko & Chang (2021) – Show how integrating investor sentiment from news and forums via BERT + LSTM-CNN significantly boosts directional prediction performance.
• Latif et al. (2023) – Demonstrate that macroeconomic indicators (VIX, EPU, FSI, shadow rates) can outperform technical indicators in forecasting S&P 500 returns using deep learning models.

---

🧪 Research Hypotheses

We test the hypothesis that machine learning models, especially those using both technical and macroeconomic data, can outperform random guessing in predicting daily returns or market direction of European stock indices.

We will compare:

• Regression models: to forecast the exact daily return
• Classification models: to predict the direction (up/down)

---

🗃️ Data Sources

The project uses publicly available and easily accessible datasets:

• 📈 Historical market data via yfinance – OHLCV prices for CAC 40, EURO STOXX 50, and STOXX Europe 600.
• 📊 Technical indicators – Locally computed (SMA, EMA, RSI, MACD, etc.).
• 🌍 Macroeconomic variables – From FRED, World Bank, etc. (e.g., VIX, interest rates, oil prices, policy uncertainty).
• 💬 (Optional) Sentiment data from Kaggle datasets or news APIs.