Skoltech NLA project 2019:

Feature Ranking via Eigenvector Centrality

Team 35 members:

• Aleksandr Belov [Aleksandr.Belov@skoltech.ru]
• Ivan Baybuza [Ivan.Baybuza@skoltech.ru]
• Mariia Kopylova [Mariia.Kopylova@skoltech.ru]
• Miron Kuznetsov [Miron.Kuznetsov@skoltech.ru]

Abstract

The problem of features ranking arises in many practical fields. In this paper, we work with medical data to rank the potential predictors of medical issue called «fibrillation». To do it, we use an Eigenvector Centrality (EC) method, proposed by the work Ranking to Learn: Feature Ranking and Selection via Eigenvector Centrality (2017). The result of this will be delivered to the National Cardiology Center in Moscow and could be reasonable for the future real practice.

References

1. Giorgio Roffo, Simone Melzi, "Ranking to Learn: Feature Ranking and Selection via Eigenvector Centrality", 2017.
2. Noah Lee, Andrew F. Laine, Jianying Hu, Fei Wang, Jimeng Sun, Shahram Ebadollahi, "Mining electronic medical records to explore the linkage between healthcare resource utilization and disease severity in diabetic patients", 2011.
3. Cheng Guo, Felix Berkhahn, "Entity Embeddings of Categorical Variables", 2016.