欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对已有的情感生理参数样本类内聚合度低、不同状态较难区分的特点,提出了一种改进的模糊支持向量机识别方法。模糊隶属度函数采用高斯分布形式,高斯分布的参数分别由同类样本数据形成的最小超球体半径和样本之间的紧密程度决定。该方法计算样本模糊隶属度时,不仅考虑样本与类中心的距离关系,还要考虑样本与样本之间的关系。实验显示改进的模糊支持向量机方法识别性能得到提高。
英文摘要：
Due to low similarity for the same emotional state parameters and difficult to distinguish between different emotional states,this paper proposed an improved fuzzy support vector machine recognition method.Fuzzy membership function took the form of the Gaussian function,determined Gaussian function parameters by the radius of the same state sample data smallest hyper sphere and tightness of sample data.Determined fuzzy membership value of sample by not only considered the distance between the sample classes and the center of the sample the class,but also considered the relationship betwwen samples.Experiments show that the improved fuzzy support vector machine recognition performance is improved.
同期刊论文项目
同项目期刊论文
期刊信息
位置：