Time Series Similarity Search Based on Binary Compression
Hongyuan Zheng · Computer and Digital Engineering · 2013
This paper presents a method based on binary compression(BC) to solve the problem of time series similarity search.This method is to divide time series into segments,extract some key points of each segment and transform them into a bit stream,segment is divided into sub-segment and take the middle points of sub-segments as key points.Record bit stream to achieve the purpose of compression.In this paper the method prove that is no false dismissals.Then use the skyline on time series index.Experimental results show that,the method based on BC method has smaller borders PAA,in similarity search the method BC with skyline index is more efficient then the method PPA based on R* tree.