Proposta e Implementação de métodos para a extração de padrões sequenciais frequentes na biblioteca PyMove com a utilização do algoritmo Prefixspan

Abstract

With technological evolution, large volumes of data are collected and stored quickly and practically on a daily. Thus, it is possible to obtain data of the most varied types, including in relation to the movements of objects, such as humans, animals, or vehicles, through geographic space. The extraction of useful information from these data, considered trajectories, can result in contributions of significant value to emerging research, such as social applications in academic and practical environments of society. However, knowledge extraction is a challenge, especially in relation to the identification of frequent behaviors in these data. Thus, this work presents a method for the extraction and treatment of frequent sequential patterns in trajectory data. The PrefixSpan algorithm was applied to an anonymized dataset, considering three proposed approaches to extract frequent sequential patterns. These patterns make it possible to infer behavior profiles of a given object through the repetition of its movements, in addition to identifying the places most visited by that object.