Article ID Journal Published Year Pages File Type
409085 Neurocomputing 2008 9 Pages PDF
Abstract

Through microarray technology a large amount of gene expression data is available for analysis. Different approaches to modeling have been proposed. It has been observed that genetic regulatory networks share many characteristics with Boolean networks such as periodicity, self-organization, etc. Moreover, it is also a known fact that in these networks, most genes are governed by Canalizing Boolean functions. However, the actual gene expression level measurements are continuous valued. To combine discrete and continuous aspects, Zhegalkin Polynomials can be used as continuous representations of Boolean functions. The requirement for the Boolean function to be canalizing can be extended to continuous functions by demanding monotonicity with respect to at least the canalizing variable. In this paper it is proven that Canalizing Zhegalkin Polynomials observe this monotonicity property. Moreover, for correct handling of normalized data it is shown that the value of a Zhegalkin Polynomial also lies within the unit interval as long as the values of its input variables also do so.

Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, , ,