论文部分内容阅读
目前,大数据问题亟待解决,关键就是对问题的特征描述.目前特征描述最流行的理论是深度学习理论,但深层结构共需要多少层,每层需要多少特征?这是深度学习最需要解决的问题.引入商空间理论对深度学习理论进行改进,根据粒度变换原理对问题特征进行深层表示,克服深度学习理论中深度不确定,特征描述不明确的缺点.首先根据商空间理论的粒度变换原则,在多粒度空间分层描述问题特征,从而形成多层的深度特征表示.接着,根据商空间粒度变换的描述特性,在不同粒度空间对问题进行求解.最后,作者选取Letter-recognition数据集进行实验,实验结果表明本文所提的深度特征表示法可以自动将问题分为多层结构,分层描述问题的特征,提升了问题求解精度.
At present, the problem of big data needs to be solved urgently, and the key point is to describe the characteristics of the problem.At present, the most popular theory of feature description is deep learning theory, but how many layers are required for deep structure and how many features are needed for each layer? This is the most needed solution for deep learning Problem.This paper introduces the quotient space theory to improve the deep learning theory and deeply expresses the question features according to the principle of granularity transformation to overcome the shortcoming of the deep indefiniteness and ambiguous description of the features in the deep learning theory.First, according to the granularity transformation principle of quotient space theory, The multi-granularity space is used to describe the features of the problem hierarchically, so as to form a multi-layered depth feature representation.Next, the problem is solved in different granular spaces according to the describing characteristics of the quotient space granularity transformation.Finally, the author selects the Letter-recognition data set for experiments The experimental results show that the depth feature representation proposed in this paper can automatically divide the problem into multi-layer structure and describe the characteristics of the problem in a hierarchical manner, which improves the accuracy of problem solving.