Attachments forums

List of attachments posted on this forum.


All files on forums: 134177

seekers, Thu Mar 02, 2017 12:04 am

We propose a forecasting procedure based on multivariate dynamic kernels, with the capability of integrating information measured at di?erent frequencies and at irregular time intervals in ?nancial markets. A data compression process rede?nes the original ?nancial time series into temporal data blocks, analyzing the temporal information of multiple time intervals. The analysis is done through multivariate dynamic kernels within support vector regression. We also propose two kernels for ?nancial time series that are computationally e?cient without a sacri?ce on accuracy. The e?cacy of the methodology is demonstratedby empirical experiments on forecasting the challenging  S&P500 market
All files in topic