Eficiencia de la inteligencia artificial como modelos de optimización y pronóstico en los mercados bursátiles: Efficiency of artificial intelligence as optimization and forecasting models in stock markets

The document shows an analysis carried out through Artificial Intelligence models to make comparisons in the predictions of investment portfolios in Mexico and the United States. ARIMA models were used for the prediction of time series in comparison with the multilayer perceptron neural network and particle swarm in the optimal portfolio selection. Markowitz's theory (1952) is also applied to identify the type of portfolio on the efficient frontier. The study shows that artificial intelligence can help in the design of better investment strategies in a given time. The document is limited by the fact that only the main companies from the aforementioned countries were taken, but the model can be replicated with larger samples even in other stock exchanges. Key words: Investment portfolios, Investment strategies, Artificial intelligence. Jel codes: G11, G12, G14