Abstract: Audio rating is one of the most important indicators of TV industry, which is of great value to decision making of TV station. According to characters of more impact factors and complexity of variety, a new method of hyper-sphere Support Vector Machine（SVM） multi-class classification based on semi-fuzzy kernel clustering is proposed. The new method defines confusion classes based on semi-fuzzy kernel clustering and sphere support vector machines using the information from boundary so as to improve the performance of classifier efficiently. Experimental results indicate that the new method yields higher precision and speed than classical classification methods.

Key words: audio rating, semi-fuzzy cluster, Support Vector Machine（SVM）

摘要： 收视率是电视行业重要的指标之一，对电视机构运营决策具有重要参考价值。针对收视率数据影响因素众多，变化趋势复杂等特点，提出了一种基于半模糊核聚类的超球支持向量机分类方法，基于半模糊核聚类生成模糊类，在其边缘样本信息基础上，利用超球支持向量机进行多类分类，从而有效提高分类器性能。实验表明，该方法比传统方法具有更高的速度和精度。

关键词: 收视率, 半模糊聚类, 支持向量机