In the past few years, due to their ability to extract multivariate correlations, machine learning tools have become more and more important for discovery of information in very complex data sets. This has had specific application to various data sets related to human brain tasks. However, this is far from a simple and direct methodology. Some of the issues involve dealing with the extreme signal to noise ratios, as well as variation between different individuals. Moreover, the huge amount of features relative to the number of data points is a challenge. As a result, in attacking these problems, we found it necessary to adapt a large variety of methodologies; chosen to overcome specific obstructions for specific problems. In this paper, we describe our experience working on several examples at the edge of capabilities of these systems and describe the various and variant methodologies we needed to overcome these sort of challenges. Hopefully these cases will serve as a guideline for other applications.
Bibliographical noteFunding Information:
This work was partially supported by a grant for computational equipment by the Caesarea Rothschild Institute at the University of Haifa, and by a Hardware Grant by NVIDIA Corporation The paper was partially written during a sabbatical visit of L. Manevitz graciously hosted at the Computer Science Department, Otago University, Dunedin, New Zealand. Some of this work has appeared in the Ph.D. thesis of Alex Frid supervised by Larry Manevitz.
© 2019, Springer Nature Switzerland AG.
- Analysis of cognitive processes
- Brain correlates
- Feature selection
- Machine learning
ASJC Scopus subject areas
- Artificial Intelligence
- Applied Mathematics