RELATE: Preference-aware Correlation-based Query Refinement

Abstract

Data exploration techniques aim to efficiently and effectively guide users towards interesting data within massive, complex data. Being one of those techniques, query refinement goal is to automatically refine a user's query so that its result satisfies a certain constraint. In this paper, we focus on time series correlation as a constraint for refinement such that the result of a query is a pair of two subsequence time series. Differently than most existing works, we make use of the deviation from a correlation constraint as an objective to minimize in our problem. Moreover, we include users preferences as an objective to maximize in proportion to users input queries. The combination of these two objectives are prevalent in applications where users seek time series subsequences that exhibit a certain correlation value range and are close from their initial queries. Towards finding the optimal query satisfying these objectives, we propose an efficient computational-centric algorithm and an innovative I/O-centric algorithm as well. We experimentally validate the efficiency and effectiveness of our proposed algorithms using real and synthetic data compared to recently proposed correlation-based subsequence search algorithms.