Sparse Binary Optimization for Text Classification via Frank Wolfe Algorithm

Text classification is a fundamental problem and has attracted significant research attention, especially in large scale search with high dimensions. We propose a novel framework for text classification, combining with an efficient sparse binary optimization algorithm called Frank Wolfe based optimization algorithm. Specifically, we project the input data into a higher-dimensional space while ensuring the sparsity of the data in the output space, and the objective function is preserving pairwise similarities between the input sample. Experimental results on benchmark datasets demonstrate the effectiveness of the proposed framework with the optimization algorithm.