Abstract:The similarity search on sensor data generated by a myriad of sensing devices is a frequently encountered problem in the era of the Internet of Things (IoT). This sensor data generally appear in the form of time series, a temporally ordered sequence of real numbers obtained regularly in time. It has been widely accepted that the dynamic time warping (DTW) currently is the most prevalent similarity measure in the time-series mining community, mainly due to its flexibility and broad applicability. However, calculating DTW between two time series has quadratic time complexity, leading to unsatisfactory efficiency when performing the similarity search over the large time-series data set. The main contribution of this article is to propose a method called product quantization (PQ)-based DTW (PQDTW) for fast time-series approximate similarity search under DTW. The PQ, a well-known approximate nearest neighbor search approach, is used in PQDTW. Nevertheless, the conventional PQ is developed with the Euclidean distance and is not designed for DTW. To solve this problem, the DTW barycenter averaging (DBA) technique is utilized to adapt the PQ for DTW before using it. We employ PQDTW along with the filter-and-refine framework to efficiently and accurately perform the time-series similarity search. Our method can reasonably reduce many DTW computations in the filtering phase; thus, the query process is accelerated. We compare PQDTW with related popular algorithms using public time-series data sets. Experimental results verify that the proposal achieves the best tradeoff between query efficiency and retrieval accuracy compared to the competitors.

Time Series Similar Pattern Matching Based on Singularity Event Features

A Time Series Similar Pattern Matching Algorithm Based on Singularity Event Features

Dynamic Time Warping under Product Quantization, with Applications to Time-Series Data Similarity Search

A Novel Similarity Measure Approach for Time Series Based on PLA and DTW

Singularity Detection and Feature Extraction Based on Analytic Wavelet Transform

A Novel Method for Time Series Symbolization Based on Singular Event Features Clustering

Time Series Symbolization Based on Singular Event Feature Clustering

Online Series Pattern Detection Based on Advanced Segmental Semi-Markov Model

Trend feature extraction method based on important points in time series

Integrating Data-Driven Segmentation, Local Feature Extraction and Fisher Kernel Encoding to Improve Time Series Classification

Segmental semi-markov model based online series pattern detection under arbitrary time scaling

The Comparison of Time-frequency Analysis Methods and Their Applications

A Hybrid Online Series Pattern Detection Algorithm

Study on Pattern Similarity of Time Series

A New Method for Similarity Matching of Non-Stationary Time Series Based on Fractal Time-Varying Dimension

Sequence Pattern Matching over Time-Series Data with Temporal Uncertainty

Multi-scale Harmonic Mean Time Surfaces for Event-based Object Classification

Efficient Algorithm for a Novel Pattern of Time Series

An Advanced Segmental Semi-Markov Model Based Online Series Pattern Detection

Matching Heterogeneous Events with Patterns

Learning the feature distribution similarities for online time series anomaly detection