基于最邻近模型的知识库多层次标签分类算法

Multi⁃level label classification algorithm for knowledge base based on nearest neighbor model

传统的知识库多层次标签分类算法分类精度低,为了解决这一问题,基于最临近模型的知识库研究一种新的多层次标签分类算法。该算法对知识库多层次标签进行特征提取,将提取的对象标签特征进行特征降维,以此获得简化知识库多层次的特征空间,通过最临近模型计算简化知识库空间中对象特征之间的距离,通过计算结果完成多层次标签分类,实现基于最临近模型的知识库多层次标签分类算法研究。实验结果表明,基于最邻近模型的知识库多层次标签分类算法能够在短时间内实现精准分类。

In order to solve this problem,a new multi⁃level label classification algorithm based on the nearest model is studied.In this algorithm,the multi⁃level label of knowledge base is extracted,and the extracted object label features are dimensionally reduced to obtain the multi⁃level feature space of the simplified knowledge base.The distance between the object features in the simplified knowledge base space is calculated by the nearest model,and the multi⁃level label classification is completed by the calculation results,and the multi⁃level label classification of knowledge base based on the nearest model is realized Class algorithm research.The experimental results show that the multi⁃level label classification algorithm based on the nearest neighbor model can achieve accurate classification in a short time.