Constructing VEGGE: Machine learning for context-sensitive graph grammars

Context-sensitive graph grammar construction tools have been used to develop and study interesting languages. However, the high dimensionality of graph grammars result in costly effort for their construction and maintenance. Additionally, they are often error prone. These costs limit the research potential for studying the growing graph based data in many fields. As interest in applications for natural languages and data mining has increased, the machine learning of graph grammars poses a prime area of research. A unified graph grammar construction, parsing, and inference tool is proposed. Existing technologies can provide a context-free tool. However, a general context-sensitive tool has been elusive. Using existing technologies for graph grammars, a tool for the construction and parsing of context-sensitive graph grammars is combined with a tool for inducing context-free graph grammars. The system is extended with novel work to infer context-sensitive graph grammars.