Graph kernels based on relevant patterns and cycle information for chemoinformatics

Chemoinformatics aim to predict molecule's properties through informational methods. Computer science's research fields concerned with chemoinformatics are machine learning and graph theory. From this point of view, graph kernels provide a nice framework for combining these two fields. We present in this paper two contributions to this research field: a graph kernel based on an optimal linear combination of kernels applied to acyclic patterns and a new kernel on the cyclic system of two graphs. These two extensions are validated on two chemoinformatics datasets.