Analysis and modelling of CMOS Gm-C filters through machine learning

Utilization of machine learning in electronics and in computer-aided design is in progress, giving an opportunity the electronic circuits to be studied in a new way that also contributes to automation of some engineering tasks. In this paper, a novel methodology for analysis and design of Gm-C filters is presented. It is based on applying classification machine learning algorithm Random Forest on theoretically gathered data sets and on published scientific results. The tree-based algorithm is chosen, because of its capability not only to identify the correct class for every training sample and to point out the decision, but also to give explanation related to this decision and to outline a set of rules. The proposed methodology is verified through creation of several data models. Gm-C filters are chosen for exploration because of their extensive usage in computer and communication systems, medical devices and sensors.