Cubic Graph Property Dataset for Machine Learning Models

Proper edge 3-coloring of a cubic graph is an NP-complete problem, which is present in number of real-live problems − correct and efficient assigning of variables used in the program to registers of the system, scheduling of a set of tasks to a set of processors, or frequency assignment of radio stations without interference. Since the finding of a solution to this problem is computationally intensive, the implementation of approaches using machine learning models for the estimation of the edge colorability is desirable. This study presents a sizable Cubic Graph Property Dataset, a graph generator focused on cubic graphs and a set of sub-datasets fitting for machine learning, deep learning and anomaly detection models.