Visualization of Repeated Patterns in Multivariate Discrete Sequences

The availability and affordability of mobile devices, wearables, sensors, IoT devices and electronic social networks produce big data in the form of complex systems of multivariate discrete sequences such as bio-informatics, natural language processing, social network corpus, etc. or non-discrete time series such as weather data, network traffic, workout data, etc. At the same time, the increased availability of advanced hardware in the form of powerful computers or high performance clusters provide us with the opportunity to analyze the aforementioned datasets that could produce vast amounts of results in diverse forms. One problem that has recently got focus is the one of discovering all repeated patterns in multivariate sequences. Novel algorithms have appeared such as ARPaD that address the specific problem however there are still no appropriate visualization methods to represent the complex results of the algorithm. In this paper, we attempt to create a visualization method that presents the common repeated patterns in multivariate discrete sequences. The visualization algorithm has been applied in a dataset of different text sequences of varying length and the results are presented in two novel type of plots, the Pattern Positional Alignment (PaPA) plot and the Stacked PaPA plot.