Abstract

We propose an approximation algorithm for efficient correlation search intime series data. In our method, we use Fourier transform and neural network toembed time series into a low-dimensional Euclidean space. The given space islearned such that time series correlation can be effectively approximated fromEuclidean distance between corresponding embedded vectors. Therefore, searchfor correlated time series can be done using an index in the embedding spacefor efficient nearest neighbor search. Our theoretical analysis illustratesthat our method's accuracy can be guaranteed under certain regularityconditions. We further conduct experiments on real-world datasets and theresults show that our method indeed outperforms the baseline solution. Inparticular, for approximation of correlation, our method reduces theapproximation loss by a half in most test cases compared to the baselinesolution. For top-$k$ highest correlation search, our method improves theprecision from 5\% to 20\% while the query time is similar to the baselineapproach query time.

 

Introduction (beta)

None

 

Conclusion (beta)

None