Two new graphs kernels in chemoinformatics

Abstract

Chemoinformatics is a well established research field concerned with the discovery of molecule's properties through informational techniques. Computer science's research fields mainly concerned by chemoinformatics are machine learning and graph theory. From this point of view, graph kernels provide a nice framework combining machine learning and graph theory techniques. Such kernels prove their efficiency on several chemoinformatics problems and this paper presents two new graph kernels applied to regression and classification problems. The first kernel is based on the notion of edit distance while the second is based on subtrees enumeration. The design of this last kernel is based on a variable selection step in order to obtain kernels defined on parsimonious sets of patterns. Performances of both kernels are investigated through experiments.

A Comparison of Explicit and Implicit Graph Embedding Methods for Pattern Recognition

“…Treelet kernel [5] is a graph kernel based on a bag of non linear patterns which computes an explicit distribution of each pattern within a graph. This method explicitly enumerates the set of treelets included within a graph.…”

“…The set of treelets, denoted T , is defined as the 14 trees having a size lower than or equals to 6 nodes. Thanks to the limited number of different patterns encoding treelets, an efficient enumeration of the number of occurrences of each labeled pattern within a graph can be computed by algorithm defined in [5]. Labeling information included within treelets is encoded by a canonical key which is defined such as if treelets have a same structure, their canonical key is similar if and only if the two treelets are isomorphic.…”

“…Note that this vector representation may be of very high dimension since it may encode all possible treelets according to all possible nodes and edges labellings defined for a graph family. In chemoinformatics, such a vector representation may have a dimension higher than 4.25 × 10 9 [5] which forbids its explicit vector embedding. Treelet kernel between graphs is defined as a sum of sub kernels between common treelets of both graphs:…”

“…Similarly to the previous study described in [6], in this paper we propose a comparison between two implicit graph embedding methods based on graph kernels ( [1,5]) and three methods of explicit graph embedding with comparable behavior ( [7,12,17]). The difference between these techniques will be illustrated on classification problems using chemoinformatic datasets, such as those from IAM [14], the predictive toxicology challenge (PTC) dataset [20] and the MAO dataset from GREYC [5].…”

“…The difference between these techniques will be illustrated on classification problems using chemoinformatic datasets, such as those from IAM [14], the predictive toxicology challenge (PTC) dataset [20] and the MAO dataset from GREYC [5]. In Section 2 and Section 3 we describe the respective graph kernels and explicit graph embedding methods.…”