Research Projects

The objective is to approach this problem through nonlinear probability modeling, where prior knowledge is required to define a sensible state representation, together with parametric forms of transition and observation probabilities. In this research, the aim is to introduce a new approach to learning nonlinear dynamic systems and show that it performs well on rather high-dimensional time series datasets compared to standard models such as Hidden Markov Models or linear predictors. The developed prediction model (deliverable) will thus help to adopt a proper preventive, restoration and utilization measure of natural resources and can also be proposed on case-specific basis which would lead to sustainable management of the pristine natural ecosystem.