Layered convolutional dictionary learning for sparse coding itemsets

Dictionary learning for sparse coding has been successfully used in different domains, however, has never been employed for the interesting itemset mining. In this paper, we formulate an optimization problem for extracting a sparse representation of itemsets and show that the discrete nature of itemsets makes it NP-hard. An efficient approximation algorithm is presented which greedily solves maximum set cover to reduce overall compression loss. Furthermore, we incorporate our sparse representation algorithm into a layered convolutional model to learn nonredundant dictionary items. Following the intuition of deep learning, our convolutional dictionary learning approach convolves learned dictionary items and discovers statistically dependent patterns using chi-square in a hierarchical fashion; each layer having more abstract and compressed dictionary than the previous. An extensive empirical validation is performed on thirteen datasets, showing better interpretability and semantic coherence of our approach than two existing state-of-the-art methods.