M-tree as an index structure for time series data

A very effective method for fast retrieval in time series data is to map each data sequence into a small set of multidimensional rectangles in feature space, then these rectangles can be readily indexed using traditional multidimensional index trees, like R∗-tree. This paper describes the use of another multidimensional index tree, M-tree, for fast retrieval in wavelet transformed time series and compares its performance to that of R∗-tree as an index structure for this transformed time series representation.