Current students

Spatial statistics models proximity-based relationships, but complex domains and structured data challenge traditional approaches. This research proposes a new mathematical framework to perform statistical analysis in network domains. Moreover, it aims to incorporate network flows and structured data, using a stochastic integral with an object-oriented approach. Specifically, a convolution process will be defined, designed to include a Markov chain dynamics. This advance will not only extend the convolutional approach to general forms of network, but it will also provide a general methodology for a geostatistical analysis in dynamical systems. ; Following the definition of the new stochastic process, research will focus on the extension and adaptations of this framework to the object-oriented approach. Creating models able to deal with these distinct features, namely domain and data complexity, can improve the comprehension of the real-world scenario, capturing more variability and providing a better fit to the data. This methodology can be applied to domains such as urban areas, transportation networks, and water systems, enabling tasks like prediction, scenario simulation, and anomaly detection.