A Novel Method for Similarity Search over Meteorological Time Series Data based on the Coulomb's Law

Several areas of knowledge use systematic and controlled observation, obtained from measurements taken at regular intervals, as a tool for behavioral analysis of phenomena, such as meteorology, which uses the observations to predict the climate behavior. Furthermore, with the advance of technology, the instruments used to measure observations have grown dramatically and the amount of data available for analysis has become greater than the ability to analyze them. In this context, this paper aims to propose a method, based on the principle of Coulomb's Law, for similarity search in time series and thus discovering intrinsic knowledge from these data. Experimental results conducted on climatic data of Brazilian cities and the sea surface temperature showed that the proposed method outperforms traditional methods on performance and accuracy and it is promising for finding similarity in series.