View the Project on GitHub ITMO-NSS-team/FEDOT.Docs

Real-world cases

We decided to prepare a set of real-world cases, that can be solved in an effective way using FEDOT.

The first case is the credit scoring problem (classification task). The task is to predict the default event for the bank customer according to the application data. The dataset is obtained from the Kaggle competition.

The second case is the metocean forecasting problem (multi-scale random process regression task). The task is to create the N-step forecast for sea surface height time series using historical values of the process. The dataset is obtained from numerical simulation of the Arctic ocean using the NEMO-LIM3 model setup (see the paper for details).

The third case is the task from the oil chemistry field: the prediction of the stability of residual fuels for sedimentation during storage in tanks.

Also, the minor cases for robotics and economics are presented.

The results of different benchmarks evaluation are also available/

The other cases are under development now.