To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
A Heterogeneous Ensemble Network Using Machine Learning Techniques |
|
Author |
Tarik Rashid |
| Citation |
Vol. 9 No. 8 pp. 335-339 |
|
Abstract |
The process of the ensemble network can be defined as grouping several networks where each independent network from that group is trained independently and then outputs are combined to obtain the overall output. This is called ensemble network. Usually, the output of ensemble network is often more accurate than outputs of any independent network. The ensemble network can be either homogenous or heterogeneous. The homogenous ensemble network obtains the overall output from different independent network structures of the same type, whereas the heterogeneous ensemble network obtains the overall output from identical independent network structures, but each independent network is trained with different training sets. In this paper, we introduce a heterogeneous ensemble network that can combine the outputs of several networks of different types. Heterogeneous ensemble network is created from recurrent neural network namely is called multi context recurrent neural network, variants of this network is used with variants functions of support vector machines to improve accuracy of the prediction task. The ensemble network is applied here to solve the problem of forecasting for electricity load energy. The network shows the ability to avoid an under or over estimated prediction of electricity load. |
|
Keywords |
Ensemble Network, Recurrent Neural Networks, Support Vector Machines, Prediction |
|
URL |
