The frequent connected subgraph mining problem, i.e., the problem
of listing all connected graphs that are subgraph isomorphic to at least
a certain number of transaction graphs of a database, cannot be solved
in output polynomial time in the general case. If, however, the transaction
graphs are restricted to forests then the problem becomes tractable.
In this paper we generalize the positive result on forests to graphs of
bounded treewidth. In particular, we show that for this class of transaction
graphs, frequent connected subgraphs can be listed in incremental
polynomial time. Since subgraph isomorphism remains NP-complete for
bounded treewidth graphs, the positive complexity result of this paper
shows that efficient frequent pattern mining is possible even for computationally
hard pattern matching operators.