提出一种基于模式聚类和混合模型参数自动选择的图库索引方法。因为传统的EM(Expectation Maximization)算法为混合模型聚类问题中的参数估计提供了一个很好的解决方法,但需要事先指定聚类数,影响了高维数据索引的精度和效率。综合利用...提出一种基于模式聚类和混合模型参数自动选择的图库索引方法。因为传统的EM(Expectation Maximization)算法为混合模型聚类问题中的参数估计提供了一个很好的解决方法,但需要事先指定聚类数,影响了高维数据索引的精度和效率。综合利用改进的CEM2(Component-wise EM of Mixture)混合模型自动选择算法、矢量量化和概率近似的索引机制,在保证准确率同时有效提高了检索效率。展开更多
Although, researchers in the ATC field have done a wide range of work based on SVM, almost all existing approaches utilize an empirical model of selection algorithms. Their attempts to model automatic selection in pra...Although, researchers in the ATC field have done a wide range of work based on SVM, almost all existing approaches utilize an empirical model of selection algorithms. Their attempts to model automatic selection in practical, large-scale, text classification systems have been limited. In this paper, we propose a new model selection algorithm that utilizes the DDAG learning architecture. This architecture derives a new large-scale text classifier with very good performance. Experimental results show that the proposed algorithm has good efficiency and the necessary generalization capability while handling large-scale multi-class text classification tasks.展开更多
Apparent biases in decision making by animals, including humans, seem to present an evolutionary puzzle, since one would expect decisions based on biased （unrealistic） information to be suboptimal. Although cognitiv...Apparent biases in decision making by animals, including humans, seem to present an evolutionary puzzle, since one would expect decisions based on biased （unrealistic） information to be suboptimal. Although cognitive biases are hard to diag- nose in real animals （Marshall et al., 2013b）, we investigate Trivers＇ proposal that individuals should self-deceive first in order to better deceive others （Trivers, 2011）. Although this proposal has been scrutinized extensively （Bandura et al., 2011） it has not been formally modelled. We present the first model designed to investigate Trivers＇ proposal. We introduce an extension to a re- cent model of the evolution of self-deception （Johnson and Fowler, 2011）. In the extended model individuals make decisions by taking directly into account the benefits and costs of each outcome and by choosing the course of action that can be estimated as the best with the information available. It is shown that in certain circumstances self-deceiving decision-makers are the most evolutionarily successful, even when there is no deception between these. In a further extension of this model individuals addi- tionally exhibit deception biases and Trivers＇ premise （that effective deception is less physiologically costly with the aid of self-deception） is incorporated. It is shown that under Trivers＇ hypothesis natural selection favors individuals that self-deceive as they deceive others .展开更多
文摘提出一种基于模式聚类和混合模型参数自动选择的图库索引方法。因为传统的EM(Expectation Maximization)算法为混合模型聚类问题中的参数估计提供了一个很好的解决方法,但需要事先指定聚类数,影响了高维数据索引的精度和效率。综合利用改进的CEM2(Component-wise EM of Mixture)混合模型自动选择算法、矢量量化和概率近似的索引机制,在保证准确率同时有效提高了检索效率。
文摘Although, researchers in the ATC field have done a wide range of work based on SVM, almost all existing approaches utilize an empirical model of selection algorithms. Their attempts to model automatic selection in practical, large-scale, text classification systems have been limited. In this paper, we propose a new model selection algorithm that utilizes the DDAG learning architecture. This architecture derives a new large-scale text classifier with very good performance. Experimental results show that the proposed algorithm has good efficiency and the necessary generalization capability while handling large-scale multi-class text classification tasks.
文摘Apparent biases in decision making by animals, including humans, seem to present an evolutionary puzzle, since one would expect decisions based on biased （unrealistic） information to be suboptimal. Although cognitive biases are hard to diag- nose in real animals （Marshall et al., 2013b）, we investigate Trivers＇ proposal that individuals should self-deceive first in order to better deceive others （Trivers, 2011）. Although this proposal has been scrutinized extensively （Bandura et al., 2011） it has not been formally modelled. We present the first model designed to investigate Trivers＇ proposal. We introduce an extension to a re- cent model of the evolution of self-deception （Johnson and Fowler, 2011）. In the extended model individuals make decisions by taking directly into account the benefits and costs of each outcome and by choosing the course of action that can be estimated as the best with the information available. It is shown that in certain circumstances self-deceiving decision-makers are the most evolutionarily successful, even when there is no deception between these. In a further extension of this model individuals addi- tionally exhibit deception biases and Trivers＇ premise （that effective deception is less physiologically costly with the aid of self-deception） is incorporated. It is shown that under Trivers＇ hypothesis natural selection favors individuals that self-deceive as they deceive others .
