摘要
为适应大数据应用背景下本体数据的计算和处理,越来越多的稀疏学习算法被应用于本体相似度计算和本体映射.在稀疏学习框架下,本体函数的学习归结于本体稀疏向量的学习.因此,利用分离Bregman方法得到本体稀疏向量计算策略,通过原始优化问题和对偶优化问题的交替迭代策略得到鞍点,进而得到最优本体稀疏向量,最后通过实验验证算法的有效性.
In order to adapt the computing and processing of ontology data in the background of big data applications,more and more sparse learning algorithms are applied to the ontology similarity calculation and the ontology mapping.Under the setting of sparse learn-ing,the learning of ontology function attributes to the learning of sparse vector.So we present an ontology sparse vector computing strate-gy by virtue of split Bregman methods.The saddle point is obtained in terms of iterative algorithm alternating between the primal and the dual optimization to get the optimal solution of ontology sparse vector and last,the effectiveness of the algorithm is verified by experi-ments.
出处
《昆明学院学报》
2015年第6期112-115,共4页
Journal of Kunming University
基金
国家自然科学青年基金资助项目(11401519)
