Open AccessRespiratory Motion Prediction with Random Vector Functional Link (RVFL) Based Neural Networks
Author(s) -
Asad Rasheed,
Abdulyekeen T. Adebisi,
Kalyana C. Veluvolu
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1626/1/012022
Subject(s) - computer science , artificial neural network , overfitting , maxima and minima , artificial intelligence , backpropagation , regularization (linguistics) , algorithm , machine learning , mathematics , mathematical analysis
Respiratory motion exhibits non-linear and non-stationary behavior in nature and this has been a great hindrance to the accurate prediction of tumor in motion adaptive radiotherapy. Accurate prediction of respiratory motion and subsequent tracking of tumor has been a challenge due to the irregularities and intra-trace variabilities. In order to overcome this issue, prediction models can be trained by using neural networks. However due to the burden of large training data, computational efficacy of existing neural networks can be affected. Moreover, training of neural networks using conventional methods like back-propagation (BP) may result in local minima and it may slow down the learning rate and convergence respectively. As a solution, in this paper, we employed random vector function link (RVFL) based neural networks to train the model in a very efficient way to achieve high accuracy in respiratory motion prediction. In RVFL, the direct link from input features to output layer acts as regularization to prevent the network from overfitting. The proposed method is tested with real respiratory motion traces acquired from 31 patients. Results show that RVFL with the use of direct link performs quite better than without direct link.