Coupled hidden Markov models for complex action recognition
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Experiments on the application of IOHMMs to model financial returns series
IEEE Transactions on Neural Networks
Support vector machine with adaptive parameters in financial time series forecasting
IEEE Transactions on Neural Networks
| Hi-index | 0.00 |
In the field of financial time series analysis it is widely accepted that the returns (price variations) are unpredictable in the long period [1]; nevertheless, this unappealing constraint could be somehow relaxed if sufficiently short time intervals are considered. In this paper this alternative scenario is investigated with a novel methodology, aimed at analyzing short (local) financial trends for predicting their sign (increase or decrease). This peculiar problem needs specific models --- different from standard techniques used for estimating the volatility or the returns --- able to capture the asymmetries between increase and decrease periods in the short time. This is achieved by modeling directly the signs of the local trends using two separate Hidden Markov models, one for positive and one for negative trends. The approach has been tested with different financial indexes, with encouraging results also in comparison with standard methods.