This paper presents a new algorithm to produce a near optimal mixture of experts model (MEM) architecture for a continuous mapping. The MEM is applied to a new method incorporating photon scatter for designing compensators for intensity modulated radiation therapy. The algorithm utilizes the fuzzy C-means clustering algorithm to partition data before training commences. A reduction in the size of training sets also allows the Levenberg–Marquardt algorithm to be implemented. As a result, both training time and validation error are reduced. A 71% reduction in prediction error compared with that of a single neural network is achieved.
Bibliographical noteThis paper is not yet available on the repository
FunderEngineering and Physical Sciences Research Council, Industrial Case Training award no. 02303507
- Committee machines
- Neural networks
- Fuzzy C-means
- Radiation therapy
Goodband, J. H., Haas, O. C. LK., & Mills, J. A. (2006). A mixture of experts committee machine to design compensators for intensity modulated radiation therapy. Pattern Recognition, 39(9), 1704-1714. https://doi.org/10.1016/j.patcog.2006.03.018