Approximate query evaluation using linear constraint databases

This paper shows that constraint databases can be used for the approximation of several types of discretely recorded continuous data, for example time series data and some spatio-temporal geographic data. We show that time series data can be approximated by a piecewise linear approximation that runs in linear time in the number of data points, and the piecewise linear approximation can be represented in a linear constraint database. Similarly, the spatio-temporal geographic data that is composed of a set of spatial locations, where each location is associated with a time series, can be also approximated and represented in a linear constraint database. The approximations provide data compression, faster query evaluation-that preserve high precision and recall-and interpolation enabling the evaluation of queries that could not be evaluated before