Metric Suffix Array For Large-Scale Similarity Search

We propose the Metric Suffix Array (MSA), as a novel and efficient data structure for permutation-based indexing. The Metric Suffix Array follows the same principles as the suffix array. The suffix array is mainly used for text indexing. Here, we build the MSA as an alternative for large-scale content based information retrieval. We also show how the MSA is scalable for parallel and distributed architectures. We study the performance and efficiency of our algorithms in a large-scale context. Experimental results show fast response time with high efficiency and effectiveness.