Discriminative frequent subgraph mining with optimality guarantees

Abstract

The goal of frequent subgraph mining is to detect subgraphs that frequently occur in a dataset of graphs. In classification settings, one is often interested in discovering discriminative frequent subgraphs, whose presence or absence is indicative of the class membership of a graph. In this article, we propose an approach to feature selection on frequent subgraphs, called CORK, that combines two central advantages. First, it optimizes a submodular quality criterion, which means that we can yield a near-optimal solution using greedy feature selection. Second, our submodular quality function criterion can be integrated into gSpan, the state-of-the-art tool for frequent subgraph mining, and help to prune the search space for discriminative frequent subgraphs even during frequent subgraph mining.

LC-mine: a framework for frequent subgraph mining with local consistency techniques

“…-The anti-cancer screen datasets (nci) consist in eight datasets collected from the PubChem Web site as in [22]. -The AIDS antiviral screen data (aids) contain the activity test information of 42,285 chemical compounds.…”

Deep topology classification: A new approach for massive graph classification

“…Graph kernels include random walks [12], shortest path [13] and discriminative subgraphs [14]. Readers are referred to a survey on graph kernels for a greater discussion on their uses outside of classification [5].…”

Blending domain estimates from two victimization surveys with possible bias

“…The patterns can range from the simple to the complex. Specifically the kernels are designed to exploit random walks [4,8,9,17], shortest paths [10], cyclic patterns [11], subtrees [2,3,12,13], and subgraphs [14–16]. Another class of graph kernels, e.g.…”

“…Frequent subgraph mining can also be used to define a kernel between two graphs [16]. Let ℱ = { s 1 ,…,s p } denote the set of p frequent and discriminative subgraph patterns mined from 𝒟.…”

“…There has been a lot of research activity in trying to develop more effective and efficient graph kernel functions κ . These methods can broadly be classified into methods based on random walks [8,9], shortest paths (SPs) [10], cycles [11] subtrees [2,12,13], and subgraphs [14,15,16]. Despite the research above, it is fair to say that efficient and effective graph classification still remains a challenge, especially for large graphs.…”