Regresar

Improving with Metaheuristics the Item Selection in Parallel Coordinates Plot

Abstract:

Data visualization is one of the most powerful techniques to analyze and obtain reliable results based on the displayed outputs since it allows humans to improve decision-making by visually analyzing data behavior. Nevertheless, it could be disrupted by high data amounts in the visualization, as is the case with Parallel Coordinate Plot (PCP), where data behavior and associations of volumes of data are difficult to identify. This paper aims to reduce this issue with PCP and take advantage of metaheuristics for optimization problems through a Simulated Annealing (SA) algorithm. The proposed method was developed and tested using air pollution and meteorological variables. The obtained results presented a reduction in data volume, thus helping represent the most relevant data for the final user.