learning process in general support vector machines take a lot of memory
optimizing slow
is not conducive to practical application. This paper presents a boundary vector-based SVM adjustable entropy function method. First
we use two methods convex hull boundary vectors relative pre-extracted boundary vectors; Then
to train pre-drawn boundary vectors in SVM adjustable entropy method. Experiments show that this method not only reduces the training sample set the price of learning
but also improve the classification rate.
关键词
Keywords
references
O L Mangasarian. Mathematical programming in data mining[J]. Data Mining and Konuolodge Discorery, 1997, 2(1): 183-201. DOI: 183
New York V N Vapnik. The Nature of Statistical Learning Theory[M]. New York: Springer, 2000: 25-46. DOI: 25