Machine Learning Methods with Graph-Level Features and Application Use Cases

— In this paper we talk about graph-level characteristics and graph kernels, which will enable us to forecast the behaviour of entire graphs. Therefore, the objective is to have a single feature that describes the overall graph's structure. We discussed graph kernels, specifically the graphlet kernel and the WL, which stands for the Weisfeiler-Lehman graph kernel. Its runtime scales only linearly in the number of edges of the graphs and the length of the WeisfeilerLehman graph sequence. Finally, we give some WL use cases from the literature