Contains the main implementations of programs for the paper: GP-based methods for domain adaptation: Using brain decoding across subjects as a test-case, by R. Santana, L. Marti, and M. Zhang (Submitted for publication).
This project implements different approaches to the creation of classifiers for MEG data in brain-decoding experiments. The description of the main steps for data processing and the explanation of how to run the different programs is given in the file Steps_For_Problem_Solution.pdf.
The evolutionary algorithms are based on the DEAP library that implements evolutionary algorithms (https://github.com/deap/deap). The importance weighting crossvalidation schemes which are used with the classical classifiers are implemented using the libtlda Python library https://github.com/wmkouw/libTLDA, which is a library of transfer learners and domain-adaptive classifiers.