Historical identification and forecast values of IBEX 35 & IPC financial indices using differential neural networks

One of the main objectives of quantitative analysis in Finance is to be able to make accurate forecasts of prices and yields from financial assets, the question is to find reliable methods and techniques to perform properly such forecasts. In the early 1990 ́s while pursuing better results the use of artificial neural networks (ANN) started, this technique was only used as a tool of monitoring and description of values, but later it evolved into a way of forecasting the behavior of economic and financial variables, achieving promising results. In this paper we present one of the first applications in finance of the Differential Neural Networks (DNN), we use that to carry out two processes: the description of daily closing values of financial index IBEX 35 (Madrid Stock Exchange) and the IPC index (Mexican Stock Exchange), both from January 3, 2000 to January 20, 2012. Later, we perform the forecast of daily closing values of these indices from January 23 to February 17, 2012. The results obtained in the description and forecast of the closing values of both indices were excellent, which makes it extremely attractive to continue the study of this methodology to carry out these forecasts.