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

ABSTARCT :

? Time series are generated at an unprecedented rate in domains ranging from finance, medicine to education. ? Collections composed of heterogeneous, variable-length and misaligned times series are best explored using a plethora of dynamic time warping distances. However, the computational costs of using such elastic distances result in unacceptable response times. ? We thus design the first practical solution for the efficient GENeral EXploration of time series leveraging multiple warped distances. ? GENEX pre-processes time series data in metric point-wise distance spaces, while providing bounds for the accuracy of corresponding analytics derived in non-metric warped distance spaces. ? Our empirical evaluation on 66 benchmark datasets provides a comparative study of the accuracy and response times of diverse warped distances. ? We show that GENEX is a versatileyet highly efficient solution for processing expensive-to-computewarped distances over large datasets, with response times 3 to 5orders of magnitude faster than state-of-art systems.

EXISTING SYSTEM :

? The “edit operations” we use the paradigm of a graphical editing process and end up with a dynamic programming algorithm that we call Time Warp Edit Distance (TWED). ? TWED is slightly different in form from Dynamic Time Warping, Longest Common Subsequence or Edit Distance with Real Penalty algorithms. ? In particular, it highlights a parameter which controls a kind of stiffness of the elastic measure along the time axis.

DISADVANTAGE :

? Mitigating this problem for a large array of distances at the theoretical rather than empirical level opens the door for increased versatility by allowing the in corpora-tion of new distances, while guaranteeing accurate similarity exploration results with response times up to 5 orders of magnitude faster than existing systems. ? Mitigating this problem for a large array of distances at the theoretical rather than empirical level opens the door for increased versatility by allowing the in corpora-tion of new distances, while guaranteeing accurate similarity exploration results with response times up to 5 orders of magnitude faster than existing systems.

PROPOSED SYSTEM :

? In this paper we address the case of elastic metrics, namely elastic similarity measures that jointly exploit time shifting and possess all the properties of a distance, in particular the triangle inequality. Our contribution is basically four folded: ? The first contribution of this paper is the proposal of a new elastic metric which we call TWED (“Time Warp Edit Distances”). ? The second contribution is related to the introduction of a parameter we call stiffness which drives the elasticity of TWED, placing this kind of distance in between the Euclidian distance (somehow a distance with ‘infinite stiffness’) and DTW (somehow a similarity measure with no ‘stiffness’ at all). ? The third contribution proposes a lower bound for the TWED measure which allows relating the evaluation of the matching of two time series into down sampled representation spaces to the evaluation of their matching into their original representation spaces. ? The fourth contribution of the paper is an empiric evaluation of the quality of TWED based on a simple classification experiment that provides some highlights on the effectiveness of TWED compared to the Euclidian Distance (ED), DTW, LCSS and ERP. The influence of the stiffness parameter on classification error rates is also analyzed.

ADVANTAGE :

? Our proposed GENEX technology can be used to explore large datasets with in seconds. ? High data cardinality leading to a compromise between increased responsiveness and higher accuracy. ? We instead report the online performance which reflects the analyst‘s experience . Both competitors take advantage of the well-known lower boundLBkeogh optimization in their implementations.