Λογότυπο αποθετηρίου

Πλοήγηση ανά Συγγραφέα "Charalampi, Kleopatra"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω
Φιλτράρετε τα αποτελέσματα πληκτρολογώντας τα πρώτα γράμματα
Τώρα δείχνει 1 - 1 από 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    Equity return forecasting & portfolio optimization: a machine-learning approach
    (2025-10-07) Charalampi, Kleopatra; Χαραλάμπη, Κλεοπάτρα; Psarakis, Stelios; Besbeas, Panagiotis; Vrontos, Ioannis
    In today’s dynamic financial markets, investment portfolio management is of central focus in financial research. While the portfolio selection problem is highly contingent upon reliable prediction of the future performance of stock markets, accurate forecasting of stock returns remains a great challenge to both academics and practitioners. This thesis investigates the application of machine learning (ML) and deep learning (DL) models in forecasting stock returns and constructing optimized equity portfolios. Using a subset of 25 highly liquid S&P 500 stocks, the study evaluates the predictive accuracy of Ridge Regression, eXtreme Gradient Boosting (XGBoost) and Long Short-Term Memory (LSTM). Forecasted returns are subsequently integrated into a mean-variance optimization model to select and allocate capital among the top-performing stocks at each rebalancing date. Empirical results support the superiority of advanced ML and DL models in stock return forecasting, compared to traditional penalized regression approaches, and demonstrate large economic gains to investors that incorporate them into their investment strategies.