Enhancing the Symbolic Aggregate Approximation Method Using Updated Lookup Tables

Similarity search in time series data mining is a problem that has attracted increasing attention recently. The high dimensionality and large volume of time series databases make sequential scanning inefficient to tackle this problem. There are many representation techniques that aim at reducing the dimensionality of time series so that the search can be handled faster at a lower dimensional space level. Symbolic representation is one of the promising techniques, since symbolic representation methods try to benefit from the wealth of search algorithms used in bioinformatics and text mining communities. The symbolic aggregate approximation (SAX) is one of the most competitive methods in the literature. SAX utilizes a similarity measure that is easy to compute because it is based on pre-computed distances obtained from lookup tables. In this paper we present a new similarity measure that is almost as easy to compute as the original similarity measure, but it is tighter because it uses updated lookup tables. In addition, the new similarity measure is more intuitive than the original one. We conduct several experiments which show that the new similarity measure gives better results than the original one.