Artificial Intelligence in the Contemporary Finance: Quantitative Investment Analysis of the performance of investment, risk management and investment efficiency.
0 viewsAbstract
Artificial intelligence (AI) has become one of the driving forces of contemporary finance, influencing the construction of portfolios, risk management, and trading speed and form. In this paper, I will construct a quantitative, reproducible assessment framework to determine the extent to which AI is a better (i) investment performance, (ii) volatility and risk forecast, and (iii) market efficiency in the 2015-2025 period. Based on daily data and two types of liquid multi-asset proxies (SPY, AGG, GLD, VNQ, EEM, DBC) that use a cash/risk-free proxy (BIL or DTB3), the research pits three typical AI model families in competition with simple benchmarks or linear baselines: Random Forests, gradient-boosted trees (XGBoost), and sequence models (LSTM). It is evaluated based on out-of-sample forecast accuracy (RMSE/MAE, directional accuracy), portfolio performance (annualized return, volatility, Sharpe/Sortino, maximum drawdown), and statistical inference methods typically used in finance (including Newey West corrected regressions and forecast comparison tests). The paper focuses on reproducibility, recording feature engineering decisions, validation design, sensitivity to transaction costs and regime-based robust checks. Summarizing the peer reviewed results and integrating the empirical design with the best practice, the paper has concluded that AI has the potential to add value provided that it represents non-linear interactions and is controlled with disciplined validation and monitoring, yet actual gains are cost- sensitive and require explainable conditions.
Keywords
Artificial Intelligence (AI), Blockchain Technology, Investment, Risk Management