Mining periodic patterns from spatio-temporal trajectories using FGO-based artificial neural network optimization model

Periodic patterns are occurrences that occur regularly over a long period of time at a specific location. In recent years, mining periodic patterns have become a popular area of research. There are several difficulties to map and find the periodical motion pattern of moving objects. Recently, the most challenging task is hidden periodic pattern detection from a historical object movement. To tackle these challenges, we have proposed artificial neural network (ANN) for periodic pattern mining. The proposed methodology involves three major sections namely data pre-processing, clustering, and periodic pattern mining. The two stages involved in data pre-processing are the dataset produced from the obtained matrix and the matrix form that expresses the given sub-sequence. Next, the data clustering for further process is carried out by using teaching–learning-based optimization (TLBO) algorithm. Finally, periodic pattern mining is effectively performed using artificial neural network (ANN)-based football game-based optimization (ANN-FGO) with nine constraints namely item, value, cyclic pattern, aggregate, length, sequence, similarity, duration, and gap constraints, respectively. However, the proposed model outperformed other existing methods in terms of performance.