作者: Scott Hellman , Amy McGovern , Ming Xue
DOI: 10.1109/CIDU.2012.6382191
关键词:
摘要: We introduce Ensembled Continuous Bayesian Networks (ECBN), an ensemble approach to learning salient dependence relationships and predicting values for continuous data. By training individual networks on both a subset of the data (bagging) attributes in (randomization), ECBN produces models domains that can be used identify important variables dataset between those variables. use linear Gaussian distributions within our ensembles, providing efficient network-level inference. ensembling these networks, we are able represent nonlinear relationships. empirically demonstrate outperforms meteorological forecast rainfall prediction task across United States, performs comparably results reported Random Forests.