Efficient Fuzzy Logic-Based Algorithm for Microarray Network Identification and Prediction in Bioinformatics

Fuzzy logic-based systems have been proposed as a tool for analyzing experimental data in microarrays in order to detect and predict gene regulation networks. The main drawback of this approach is the so-called curse of dimensionality associated with high dimension exhaustive searches. In this paper, it is shown that simple standard fuzzy systems can be transformed into additively separable systems by using an appropriate rotation of the input variables. By doing this, the algorithm for computing the output of the fuzzy system takes polynomial time and, therefore, it is computationally efficient for finding higher dimension interaction among genes. Ill. 3, bibl. 8 (in English; summaries in English, Russian and Lithuanian).