We are living in a world where the future is uncertain, it’s important to provide accurate predictions which is important for decision making to the community. While at Aims, Ariane worked on hierarchical approaches used to forecast times series. In her study, she discussed techniques for predicting hierarchical time series, primarily the top-down (TD) approach, the bottom-up (BU) approach and the optimal combination approach (COM). The different approaches were applied to a data set that includes the monthly number of fire spots in Brazil, between 2011 and 2020. The results show that the top-down approach produces better forecasting performances in overall hierarchy, considering the data under study.

In the same dynamics of modeling real facts using mathematical tools, Ariane is currently working on spatiotemporal models as a PhD student at Leibniz University of Hannover.