Regression based on support vector classification
ICANNGA'11 Proceedings of the 10th international conference on Adaptive and natural computing algorithms - Volume Part II
Support vector regression based on data shifting
Neurocomputing
| Hi-index | 0.00 |
The improved regression is proposed in this paper by training SVMs and only using support vectors. The method comes from the combination of support vector machines and regression. Firstly, all the training data are divided into the positive class and the negative one according to the signs of the errors. So the regression problem is transformed into the classification of twoclass problem. The support vector set is obtained by training learning machines on the training set that consists of these two-class observation data. Secondly, the proposed regression is constructed based on the obtained support vectors formerly. The results of experiments indicate that the proposed regression has smaller errors compared with the traditional regression and support vector regression.