Serena Sorrentino

NORMS: An automatic tool to perform schema label normalization

Schema matching is the problem of finding relationships among concepts across heterogeneous data sources (heterogeneous in format and structure). Schema matching systems usually exploit lexical and semantic information provided by lexical databases/thesauri to discover intra/inter semantic relationships among schema elements. However, most of them obtain poor performance on real world scenarios due to the significant presence of “non-dictionary words”. Non-dictionary words include compound nouns, abbreviations and acronyms. In this paper, we present NORMS (NORMalizer of Schemata), a tool performing schema label normalization to increase the number of comparable labels extracted from schemata1.

Automatic annotation in data integration systems

CWSD (Combined Word Sense Disambiguation) is an algorithm for the automatic annotation of structured and semi-structured data sources. Instead of being targeted to textual data sources like most of the traditional WSD algorithms, CWSD can exploit knowledge from the structure of data sources together with the lexical knowledge associated with schema elements (terms in the following).

Semantic Annotation of the CEREALAB database by the AGROVOC Linked Dataset

Abstract Nowadays, there has been an increment of open data government initiatives promoting the idea that particular data should be freely published. However, the great majority of these resources is published in an unstructured format and is typically accessed only by closed communities. Starting from these considerations, in a previous work related to a dataset on young workers on non permanent contracts, we proposed an experimental and preliminary methodology for facilitating resource providers in publishing public data into the Linked Open Data (LOD) cloud, and for helping consumers (companies and citizens) in efficiently accessing and querying them. Linked Open Data play a central role for accessing and analyzing the rapidly growing pool of life science data and, as discussed in recent meetings, it is important for data source providers themselves making their resources available as Linked Open Data. In this paper we extend and apply our methodology to the agricultural domain, i.e. to the CEREALAB database, created to store both genotypic and phenotypic data and specifically designed for plant breeding, in order to provide its publication into the LOD cloud.

Semantic Annotation and Publication of Linked Open Data

Nowadays, there has been an increment of open data government initiatives promoting the idea that particular data produced by public administrations (such as public spending, health care, education etc.) should be freely published. However, the great majority of these resources is published in an unstructured format (such as spreadsheets or CSV) and is typically accessed only by closed communities. Starting from these considerations, we propose a semi-automatic experimental methodology for facilitating resource providers in publishing public data into the Linked Open Data (LOD) cloud, and for helping consumers (companies and citizens) in efficiently accessing and querying them. We present a preliminary method for publishing, linking and semantically enriching open data by performing automatic semantic annotation of schema elements. The methodology has been applied on a set of data provided by the Research Project on Youth Precariousness, of the Modena municipality, Italy.

Automatic normalization and annotation for discovering semantic mappings

Normalization and lexical annotation methods, developed in the context of matching systems, have proven to be effective for the discovery of lexical relationships among schemata. We will show how these methods are applicable and effective in the context of Semantic Resource Framework to mine the semantics of a web service interface and to discover mappings between them.

Exploiting semantics for filtering and searching knowledge in a software development context

Software development is still considered a bottleneck for Small and Medium Enterprises (SMEs) in the advance of the Information Society. Usually, SMEs store and collect a large number of software textual documentation; these documents might be profitably used to facilitate them in using (and re-using) Software Engineering methods for systematically designing their applications, thus reducing software development cost. Specific and semantics textual filtering/search mechanisms, supporting the identification of adequate processes and practices for the enterprise needs, are fundamental in this context. To this aim, we present an automatic document retrieval method based on semantic similarity and Word Sense Disambiguation techniques. The proposal leverages on the strengths of both classic information retrieval and knowledge-based techniques, exploiting syntactical and semantic information provided by general and specific domain knowledge sources. For any SME, it is as easily and generally applicable as are the search techniques offered by common enterprise Content Management Systems. Our method was developed within the FACIT-SME European FP-7 project, whose aim is to facilitate the diffusion of Software Engineering methods and best practices among SMEs. As shown by a detailed experimental evaluation, the achieved effectiveness goes well beyond typical retrieval solutions.

Uncertainty in data integration systems: automatic generation of probabilistic relationships

This paper proposes a method for the automatic discovery of probabilistic relationships in the environment of data integration systems. Dynamic data integration systems extend the architecture of current data integration systems by modeling uncertainty at their core. Our method is based on probabilistic word sense disambiguation (PWSD), which allows to automatically lexically annotate (i.e. to perform annotation w.r.t. a thesaurus/lexical resource) the schemata of a given set of data sources to be integrated. From the annotated schemata and the relathionships defined in the thesaurus, we derived the probabilistic lexical relationships among schema elements. Lexical relationships are collected in the Probabilistic Common Thesaurus (PCT), as well as structural relationships.

An Ontology-Based Data Integration System for Data and Multimedia Sources

Data integration is the problem of combining data residing at distributed heterogeneous sources, including multimedia sources, and providing the user with a unified view of these data. Ontology based Data Integration involves the use of ontology(s) to effectively combine data and information from multiple heterogeneous sources. Ontologies, with respect to the integration of data sources, can be used for the identification and association of semantically corresponding information concepts, i.e. for the definition of semantic mappings among concepts of the information sources. MOMIS is a Data Integration System which performs information extraction and integration from both structured and semistructured data sources. In MOMIS was extended to manage “traditional” and “multimedia” data sources at the same time. STASIS is a comprehensive application suite which allows enterprises to simplify the mapping process between data schemas based on semantics. Moreover, in STASIS, a general framework to perform Ontology-driven Semantic Mapping has been pro-posed. This paper describes the early effort to combine the MOMIS and the STASIS frameworks in order to obtain an effective approach for Ontology-Based Data Integration for data and multimedia sources.

Dealing with Uncertainty in Lexical Annotation

We present ALA, a tool for the automatic lexical annotation (i.e. annotation w.r.t. a thesaurus/lexical resource) of structured and semi-structured data sources and the discovery of probabilistic lexical relationships in a data integration environment. ALA performs automatic lexical annotation through the use of probabilistic annotations, i.e. an annotation is associated to a probability value. By performing probabilistic lexical annotation, we discover probabilistic inter-sources lexical relationships among schema elements. ALA extends the lexical annotation module of the MOMIS data integration system. However, it may be applied in general in the context of schema mapping discovery, ontology merging and data integration system and it is particularly suitable for performing “on-the-fly” data integration or probabilistic ontology matching.

The CEREALAB Database: Ongoing Research and Future Challenges

The objective of the CEREALAB database is to help the breeders in choosing molecular markers associated to the most important traits. Phenotypic and genotypic data obtained from the integration of open source databases with the data obtained by the CEREALAB project are made available to the users. The first version of the CEREALAB database has been extensively used within the frame of the CEREALAB project. This paper presents the main achievements and the ongoing research related to the CEREALAB database. First, as a result of the extensive use of the CEREALAB database, several extensions and improvements to the web application user interface were introduced. Second, always derived from end-user needs, the notion of provenance was introduced and partially implemented in the context of the CEREALAB database. Third, we describe some preliminary ideas to annotate the CEREALAB database and to publish it in the Linking Open Data network.