将TLP和本体回归算法相融合,提出基于TLP经验模型的本体相似度计算和本体映射算法。新算法继承了TCP的特点,使其具有无偏参数估计的特征。将新算法应用于GO本体和物理教育本体,通过实验结果表明新算法对特定的应用领域具有较高的效率。
〔Abstract〕By combing truncated Lasso penalty (TLP) with ontology regression algorithm, this paper proposes the new ontology similarity computation and ontology mapping algorithm based on TLP empirical model. The new algorithm inherits the characteristics of TCP and has the quality of unbiased parameter estimation. The experiment shows that the new algorithm achieves higher efficiency in specific applications when it is applied to the GO and the physical education ontology.