Semantic technologies made in Evodevo!
EVODEVO was invited to participate, as Oracle partner, in the 2018 edition of the Analytics and Data Summit (BIWA Summit).
The event, dedicated to the presentation of the major innovations related to Oracle Business Intelligence, Data Warehouse, Advanced Analytics, Machine Learning and Big Data, brings together international experts in the sector every year.
On this occasion we had the pleasure of presenting our Florence Fraud Detection System project in the session on semantic technologies, developed to help the municipality of Florence in the difficult task of detecting tax evasion, thanks to the use of semantic technologies present in Oracle 12c Spatial and Graph.
The project involved the creation of an Intelligent Semantic Decision Support System with the aim of allowing automation of the aggregation process and analysis of the enormous amount of data from heterogeneous sources, assisting local institutions in the onerous task of detecting behaviours fraudulent among AIRE citizens, maximising the efficiency of the service for the benefit of the whole community.
In particular, the solution proposed by our team of experts from the Evodevo Semantic Lab for the Fraud Detection System consists, from a logical and technological point of view, in the mapping and merging of various relational databases into an Oracle 12c database performed with the Spatial & Graph option . The data stored in the various Oracle databases were transformed into triples after applying the mapping rules, according to the R2RML process compliant with the W3C recommendations.
The ontology, designed ad hoc for the domain of interest, provided the shared data structure, to combine the different data sources and a semantic structure on which the logical rules can be applied.
To complete the project, a Joseki web interface and a SPARQL endpoint supported by Oracle were created for the Fraud Detection System.
The ontologies, logic functions and rules have all been created on the basis of the international standards of the W3C.
The choice to use the data already owned by the municipality of Florence combined with the adoption of their centralized management and the inclusion of all the knowledge bases necessary for the proper conduct of the inferential processes, leads us to be able to claim to have achieved the objective of substantial support for the optimization of the identification of the most favorable conditions for evasion, within the general procedure for detecting tax fraud.