Nonlinear Causality Inference in Microarray Time Series
Detecting causal relationships between time series data has been widely studied in many areas. Granger causality that is a linear regression-based model for determining whether a single time series is useful in forecasting another; however, this approach cannot detect nonlinear relations. We use Elastic-net regularization to infer linear Granger causalities in gene expression data. Moreover, we have proposed a method which uses the properties of kernel algorithms to infer nonlinear causalities.
Fateme Nateghi received the B.Sc. degree in electrical engineering from University of Guilan, Iran, in 2011, and the M.Sc. degree in bioinformatics from Amirkabir University of Technology, Iran, in 2017. Currently, she is a PhD researcher in biomedical scineces at KU Leuven KULAK. Her current research interests is in the area of machine learning with applications in intensive healthcare.