Financial forecasting using support vector machines

The use of Support Vector Machines (SVMs) is studied in financial forecasti ng by comparing it with a multi-layer perceptron trained by the Back Propag ation (BP) algorithm. SVMs forecast better than BP based on the criteria of Normalised Mean Square Error (NMSE). Mean Absolute Error (MAE), Directiona l Symmetry (DS) Correct Up (CP) trend and Correct Down (CD) trend S&P 500 d aily price index is used as the data set. Since there is no structured way to choose the free parameters of SVMs, the generalisation error with respec t to the free parameters of SVMs is investigated in this experiment. As ill ustrated in the experiment, they have little impact on the solution. Analys is of the experimental results demonstrates that it is advantageous to appl y SVMs to forecast the financial rime series.