On improving the efficiency of complex-valued ELM
- Resource Type
- Conference
- Authors
- Chandra, B.; Sharma, Rajesh K.
- Source
- 2016 International Joint Conference on Neural Networks (IJCNN) Neural Networks (IJCNN), 2016 International Joint Conference on. :4438-4442 Jul, 2016
- Subject
- Computing and Processing
Training
Support vector machines
Testing
Performance evaluation
Feedforward neural networks
Backpropagation
Complex-valued Extreme Learning Machine
Feedforward Neural Networks
- Language
- ISSN
- 2161-4407
The Paper proposes new complex-valued ELM. The main emphasis in designing the new Complex-valued ELM is to improve the classification accuracy and also to remove the drawbacks present in other complex-valued ELMs. A novel method of converting the input to complex domain has been proposed in which the real-valued feature is considered as a projection of complex number. Two random complex valued input-hidden weights are generated which project the complex-valued input on two random hyper-planes. The ratio of two random projections of input is used as aggregation function. An algebraic function has been used as the activation function. Comparative performance evaluation of proposed complex-valued ELM with the well-known CC-ELM [1] and SVM demonstrates the superiority of proposed complex-valued ELM.