Resumen

GARCIA, Irene; TRIGO, Loren; COSTANZO, Sabatino  y  TER HORST, Enrique. Procesos gaussianos en la predicción de las fluctuaciones de la economía mexicana. El trimestre econ [online]. 2010, vol.77, n.307, pp.585-602. ISSN 2448-718X.

The ability of some neural nets to predict the direction of the Mexican economy - represented by its lei- when taking as inputs the simultaneous versions (smoothing and predictive) of a Gaussian Process fed with a Stock Index and a Bonds Index representing the Mexican market, is favorably compared - through the Anatolyev and Gerko predictive accuracy test- with the predictive ability of nets developed for a similar purpose by the authors in a previous paper, and whose inputs are the lagged indexes of the Mexican capital markets and some of their moving averages.

Palabras llave : redes neuronales; predicción; fluctuaciones; economía; mercados; mercados de capitales; procesos gaussianos.