Theoretical Characteristics on Scoring Function in Multi-dividing Setting.

The learning and optimizing of the scoring function are widely used in neural network, information retrieval and protein analysis. The multi-dividing ontology algorithms have drawn plenty of attention recent years. In this paper, we consider the Bayes-optimal scoring function in multi-dividing setting. By virtue of conditional risk, proper loss theory and derivative computing, we determine the scoring function in multi-dividing setting for certain special case. The results achieved in our paper illustrate the promising application prospects for multi-dividing ontology algorithm. [ABSTRACT FROM AUTHOR]