1 June 2016 in International Journal of Distributed Sensor Networks
Communication is a prerequisite for any form of social activity, including social networking. Nowadays, communication is not reserved only for humans, but machines can also communicate. This paper reviews the state-of-the-art technology in the area of Machine-to-Machine (M2M) communication by comparing the M2M concept with other related research paradigms such as Wireless Sensor Networks, Cyber-Physical Systems, Internet of Things, and Human-Agent Collectives. Furthermore, the paper analyses trends in the interconnecting of machines and identifies an evolutionary path in which future (smart) machines will form mostly or completely autonomous communities bonded through social connections. Such communities—machine social networks—will be formed dynamically, just like human connections, and based on the needs of machines, their context, and state of their environment. Finally, the paper outlines the current evolutionary stage and identifies key research challenges of machine social networking.
4 February 2025 in IEEE Access
One of the main challenges in ontology matching is to match ontologies with high accuracy. Therefore, ontology matching systems typically use multiple basic matchers, each targeting a specific ontology component for the matching process. However, optimizing the combination of these matchers remains an open problem. In this paper, we present CroMatcher 2.0, an improved ontology matching system that aims to overcome these challenges. We introduce two new basic matchers. The first matcher determines correspondence between entities by comparing strings obtained from entity IDs and annotations using the English lexical database to find similarities between tokens of the compared strings, considering their mutual relations (synonyms, hypernyms, etc.). The second matcher determines the correspondence between entities using the special mediator ontology, which is very valuable for ontology matching as it can contain additional information about the compared ontologies. In this paper, we tested this matcher on the Ontology Alignment Evaluation Initiative Anatomy track by using the Uberon mediator ontology, which contains a lot of information about anatomical structures. We also introduce a new weighted aggregation method (Autoweight 3.0) that automatically determines the weighting factors of the basic matchers in the parallel composition. CroMatcher 2.0 was evaluated in three test cases of the Ontology Alignment Evaluation Initiative (Benchmark, Anatomy, and Conference) and showed competitive performance compared to other state-of-the-art systems. The results position CroMatcher 2.0 among the best ontology matching systems for these datasets and confirm the effectiveness of the newly introduced methods.
26 July 2019 in Automatika
Big data have analytical potential that was hard to realize with available technologies. After new storage paradigms intended for big data such as NoSQL databases emerged, traditional systems got pushed out of the focus. The current research is focused on their reconciliation on different levels or paradigm replacement. Similarly, the emergence of NoSQL databases has started to push traditional (relational) data warehouses out of the research and even practical focus. Data warehousing is known for the strict modelling process, capturing the essence of the business processes. For that reason, a mere integration to bridge the NoSQL gap is not enough. It is necessary to deal with this issue on a higher abstraction level during the modelling phase. NoSQL databases generally lack clear, unambiguous schema, making the comprehension of their contents difficult and their integration and analysis harder. This motivated involving semantic web technologies to enrich NoSQL database contents by additional meaning and context. This paper reviews the application of semantics in data integration and data warehousing and analyses its potential in integrating NoSQL data and traditional data warehouses with some focus on document stores. Also, it gives a proposal of the future pursuit directions for the big data warehouse modelling phases.