A Datalog + Plus RuleML 1.01 Architecture for Rule-Based Data Access in Ecosystem Research

Rule-Based Data Access (RBDA) enables automated reasoning over a knowledge base (KB) as a generalized global schema for the data in local (e.g., relational or graph) databases reachable through mappings. RBDA can semantically validate, enrich, and integrate heterogeneous data sources. This paper proposes an RBDA architecture layered on Datalog RuleML, and uses it for the ∆Forest case study on the susceptibility of forests to climate change. Deliberation RuleML 1.01 was mostly motivated by Datalog customization requirements for RBDA. It includes Datalog RuleML 1.01 as a standard XML serialization of Datalog, a superlanguage of the decidable Datalog±. Datalog RuleML is customized into the three Datalog extensions Datalog[∃], Datalog[=], and Datalog[⊥] through MYNG, the RuleML Modular sYNtax confiGurator generating (Relax NG and XSD) schemas from language-feature selections. The ∆Forest case study on climate change employs data derived from three main forest monitoring networks in Switzerland. The KB includes background knowledge about the study sites and design, e.g., abundant tree species groups, pure tree stands, and statistical independence among forest plots. The KB is used to rewrite queries about, e.g., the eligible plots for studying a particular species group. The mapping rules unfold our newly designed global schema to the three given local schemas, e.g. for the grade of forest management. The RBDA/∆Forest case study has shown the usefulness of our approach to Ecosystem Research for global schema design and demonstrated how automated reasoning can become key to knowledge modeling and consolidation for complex statistical data analysis.