A lot of research effort has been directed towards applying Machine Learning (ML) methods in circuit applications, which were merely providing techniques for circuit parameter optimization like transistor length to reduce parasitic or inductance/capacitance values in power converters to minimize EMI. This paper proposes foundations on how to effectively map all physical characteristics of any electric circuit irrespective of the connection or the number of components it has to a form that can be easily interpreted and processed by machine learning algorithms. This serves as a starting point to many applications such as AI based circuit synthesis, automatic circuit layout generation, AI circuit design automation and many other applications. As a proof of concept, seven resonant circuits of increasing circuit order are mapped to ML domain, where a classifier is applied and is able to classify the circuits with an accuracy of 97.37%.