Semiparametric mixture of experts with unspecified gate network

다국어 초록 (Multilingual Abstract)

The traditional mixture of experts (ME) modeled the gate network using a certain parametric function. However, if the assumed parametric function does not properly reflect the true nature, the prediction strength of ME would become weak. For example, the parametric ME often uses logistic or multinomial logistic models for the network model. However, this could be very misleading if the true nature of the data is quite different from those models. Although, in this case, we may develop more flexible parametric models by extending the model at hand, we will never be free from such misspecification problems. In order to alleviate such weakness of the parametric ME, we propose to use the semi-parametric mixture of experts (SME) in which the gate network is estimated in a non-parametrical way. Based on this, we compared the performance of the SME with those of ME and neural networks via several simulation experiments and real data examples.

참고문헌 (Reference)

1 황선영, "이항-퇴화 혼합분포의 최우추정법" 한국데이터정보과학회 26 (26): 313-322, 2015

2 Gunther, F., "Neuralnet : Training of neural networks" 2 : 30-38, 2010

3 Young, D. S., "Mixtures of regressions with predictor-dependent mixing proportions" 54 : 2253-2266, 2010

4 Huang M., "Mixture of regression models with varying mixing proportions : A semipara-metric approach" 107 : 711-724, 2012

5 Masoudnia, S., "Mixture of experts : A literature survey" 42 : 275-293, 2014

6 Benaglia, T., "Mixtools : An R package for analyzing nite mixture models" 32 : 1-29, 2009

7 Dempster, A. P., "Maximum likelihood from incomplete data via the EM algorithm" 39 : 1-38, 1977

8 Jordan, M. I., "Hierarchical mixtures of experts and the EM algorithm" 6 : 181-214, 1994

9 이경은, "Curve Clustering in Microarray" 한국데이터정보과학회 15 (15): 575-584, 2004

10 Jacobs, R. A., "Adaptive mixtures of local experts" 3 : 79-87, 1991

11 오창혁, "A maximum likelihood estimation method for a mixture of shifted binomial distributions" 한국데이터정보과학회 25 (25): 255-261, 2014

12 Goldfeld, S. M., "A Markov model for switching regression" 1 : 3-15, 1973