PROJECT TITLE :

A General Approach For Supporting Time Series Matching using Multiple-Warped Distances

ABSTRACT:

In fields as diverse as education, medicine, and finance, new time series are being produced at a rate that has never been seen before. Exploring collections that aren't aligned, have varying lengths, and have different kinds of time series can be done most effectively with a variety of different dynamic time warping distances. However, the computational costs associated with using such elastic distances result in response times that are unacceptably slow. As a result, we come up with the first real-world solution for the effective general exploration of time series by utilizing multiple warped distances. GENEX performs preliminary processing on time series data in non-metric warped distance spaces, while simultaneously providing bounds for the accuracy of corresponding analytics derived in metric point-wise distance spaces. A comparative analysis of the accuracy and response times of various warped distances is provided by our empirical evaluation, which was performed on 66 different benchmark datasets. We demonstrate that GENEX is a flexible and highly effective solution for the processing of warped distances over large datasets, with response times that are three to five orders of magnitude faster than those of state-of-the-art systems.