稀疏字典学习方法综述

Review on Sparse Dictionary Learning Method

深度学习技术的兴起为稀疏字典学习带来新的发展契机,在大数据背景下稀疏字典学习方法有着更广阔的研究和发展空间。本文从稀疏字典学习的非监督字典学习方法和监督字典学习方法两方面进行归纳和总结,阐述非监督学习与监督学习两类稀疏字典学习方法的基本理论和特点,并对其中的典型算法进行比较、分析,将先验知识和深度学习技术加入到稀疏字典学习方法中,为稀疏字典学习方法提出新的研究思路,展望字典学习未来的研究及发展方向。

The evolution of deep learning technology has brought new development opportunities for sparse dictionary learning. The sparse dictionary learning method has a broader research and development space in the context of big data. This paper summarizes the non -supervised learning methods and supervised learning methods for sparse dictionary respectively, and expounds the basic theories and characteristics of their sparse dictionary learning methods. Then, it analyzes and compares the typical algorithms to each other. In the end, it puts forward new ideas of the sparse dictionary learning method, adding the prior knowledge and deep learning techniques to the sparse dictionary learning method is the future research and development orientation.