基于隐空间映射的联合嵌入式多标签分类算法研究

Research on Joint Embedded Multi-label Classification Algorithm Based on Hidden Space Mapping

针对当前联合嵌入式多标签分类算法,在计算过程中标签参数特征模糊,导致联合嵌入式多标签分类效果较差,分类精度较低的问题,提出了基于隐空间映射的联合嵌入式多标签分类算法.选择二分类损失焦点函数,获取多标签分类指标,根据多标签分类指标,提取联合嵌入式多标签参数特征,采用隐空间映射,筛选优化多标签参数,确保粗糙与精确结构响应方式一致,实现联合嵌入式多标签分类.实验结果表明,该算法的联合嵌入式多标签分类效果较好,能够有效提高联合嵌入式多标签分类精度.

In view of the current joint embedded multi-label classification algorithm,the label parameter features are fuzzy in the calculation process,which leads to the poor classification effect and low classification accuracy of joint embedded multi-label.Thus,a joint embedded multi-label classification algorithm based on hidden space mapping is proposed.Then,the loss focus function is selected to obtain the multi-label classification index.According to the multi label classification index,the feature of joint embedded multi label parameters is extracted,and the hidden space mapping is used to filter and optimize the multi label parameters,so as to ensure the consistency of rough and accurate structure response,and realize joint embedded multi-label classification.The experimental results show that the proposed algorithm can effectively improve the accuracy of joint embedded multilabel classification.