Multi-scale histograms for answering queries over time series data

Abstract

Similarity-based time series data retrieval has been used in many real world applications, such as stock data or weather data analysis. Two types of queries on time series data are generally studied: pattern existence queries and exact match queries. Here, we describe a technique to answer both pattern existence queries and exact match queries. A typical application that needs answers to both queries is an interactive analysis of time series data. We propose a histogram-based representation to approximate time series data.