Ontological model identification based on data from heterogeneous sources

TL;DR

This paper presents a framework for automatically creating a unified knowledge graph from heterogeneous, semi-structured data sources, improving data integration and analysis within organizations.

Contribution

It introduces a novel autonomous method for constructing knowledge graphs from diverse data sources with arbitrary structures, enhancing data consistency and reliability.

Findings

High completeness in data coverage (11/11)

Effective filtering of false connections (average 2.5 per collection)

Applicable to partially-structured data from various sources

Abstract

The development of a company often entails the emergence of autonomous data sources with different structural and technological organization. This can lead to the inability of data analysis at a high level and a violation of the integrity and reliability of data within the organization, hindering the adoption of high-quality decisions and further development of the company. This problem can be solved by implementing a higher abstraction, representing heterogeneous organization data in a single space by combining them into a single knowledge graph. We propose a framework capable of autonomous construction of an organization's knowledge graph based on semi-structured data from various sources by finding links between sources based on data with an arbitrary structure, and combining document collections into single entities. The results of tests show the applicability of the developed approach for constructing a knowledge graph based on partially-structured data from various sources and the high efficiency of the approach based on the metrics of completeness of data storage subsystems coverage (11 out of 11) and filtering false connections (there are only 2.5 connections of collections with neighbors on average in the final graph).