ADAMAP: Automatic Alignment of Relational Data Sources Using Mapping Patterns Diego Calvanese 1,2 , Avigdor Gal 3 , Naor Haba 3 , Davide Lanti 1(B ) , Marco Montali 1 , Alessandro Mosca 1 , and Roee Shraga 3 1 Free University of Bozen-Bolzano, Bolzano, Italy {calvanese,lanti,montali,mosca}@inf.unibz.it 2 Ume˚ a University, Ume˚ a, Sweden 3 Technion – Israel Institute of Technology, Haifa, Israel avigal@technion.ac.il, {naor-haba,shraga89}@campus.technion.ac.il Abstract. We propose a method for automatically extracting semantics from data sources. The availability of multiple data sources on the one hand and the lack of proper semantic documentation of such data sources on the other hand call for new strategies in integrating data sources by extracting semantics from the data source itself rather than from its documentation. In this work we focus on relational databases, observing they are created from semantically-rich designs such as ER diagrams, which are often not conveyed together with the database itself. While the relational model may be semantically-poor with respect to ontological models, the original semantically-rich design of the application domain leaves recognizable footprints that can be converted into ontology map- ping patterns. In this work, we offer an algorithm to automatically detect and map a relational schema to ontology mapping patterns and offer an empirical evaluation using two benchmark datasets. 1 Introduction Modern industrial processes and business processes require intensive use of large- scale data alignment and integration techniques to combine data from multi- ple heterogeneous data sources into meaningful and valuable information. Such integration is performed on structured and semi-structured data sets from var- ious sources such as SQL and XML schemata, entity-relationship (ER) dia- grams, ontology descriptions, process models, and web forms. Data integra- tion plays a key role in a variety of domains, including data warehouse load- ing and exchange, aligning ontologies for the Semantic Web, semantic process model matching [16], and business document format merging (e.g., orders and invoices in e-commence) [21]. As an example, consider an application that keeps track of funded project applications, managing the review process through panel meetings. One of the main challenges of data integration is to create a common seman- tic understanding from the multiple available data sources. In ontology-based c  Springer Nature Switzerland AG 2021 M. La Rosa et al. (Eds.): CAiSE 2021, LNCS 12751, pp. 193–209, 2021. https://doi.org/10.1007/978-3-030-79382-1_12