Government institutions have released a large number of datasets on their open data portals, which are in line with the data transparency and open government initiatives. With the purpose of making it more accessible and visible, these portals categorize datasets based on different criteria like publishers, categories, formats, and descriptions. However, some of this information is often missing, making it impossible to find datasets in all of these ways. As a result, with the number of datasets growing further on the portals, it is getting harder to obtain the desired information. This paper addresses this issue by introducing EODClassifier framework that suggests the best match for the category where a dataset should belong to. It relies on formal concept analysis as a means to generate a data structure that will reveal shared conceptualization originating from tags' usage and utilize it as a knowledge base to categorize uncategorized open datasets.

This article describes the low-voltage distribution network monitoring system developed for the Electric Power Industry of Serbia electricity distribution company. The implemented system is deployed in two different regional centers of the Electric Power Industry of Serbia and is being used in production for the past 4 years. During this period, it has collected real-life data which was used to analyze and investigate low-voltage distribution network performances as well as to identify and study energy losses in low-voltage distribution networks. The main contribution of this article is represented by the real-life monitoring and control infrastructure implementation insights. Furthermore, the system we describe in this article became a part of company’s enterprise information system, thus becoming integrated with heterogeneous information sources. This integration generated additional values to the company since the collected measurements started to improve the quality of the decision-making system for everyday business processes.

Producing a new high-quality text corpus is a big challenge due to the required complexity and labor expenses. High-quality datasets, considered a prerequisite for many supervised machine learning algorithms, are often only available in very limited quantities. This in turn limits the capabilities of many advanced technologies when used in a specific field of research and development. This is also the case for the Serbian language, which is considered low-resourced in digitized language resources. In this paper, we address this issue for the Serbian language through a novel approach for generating high-quality text corpora by improving text recognition accuracy for scanned documents belonging to Serbian legal heritage. Our approach integrates three different components to provide high-quality results: a BERT-based large language model built specifically for Serbian legal texts, a high-quality open-source optical character recognition (OCR) model, and a word-level similarity measure for Serbian Cyrillic developed for this research and used for generating necessary correction suggestions. This approach was evaluated manually using scanned legal documents sampled from three different epochs between the years 1970 and 2002 with more than 14,500 test cases. We demonstrate that our approach can correct up to 88% of terms inaccurately extracted by the OCR model in the case of Serbian legal texts.

Open data initiatives have resulted in a large amount of data being published on open data portals. In order to make published data more accessible these portals provide search mechanisms based on metadata like category, tags, format, and publisher. However, due to the lack of metadata information, like category values, search results are incomplete, and data is harder to be found. Therefore, this paper presents an approach for metadata enrichment on open data portals. In particular, the presented approach is focused on suggesting categories one dataset should belong to, based on the combination of tags that is used for its description. This approach relies on the tags usage for describing datasets within categories and the Formal Concept Analysis method for creating a hierarchical order of tags within categories. The presented proposal has been evaluated using data from Ireland's open data portal.

Numerous datasets have been made available on open data portals as a result of the open data initiatives. These portals offer a variety of search possibilities based on the metadata of datasets to facilitate data findability and usability. However, insufficient information frequently has a direct effect on search result quality and, in turn, data discoverability. As a result, methods for completing the missing metadata information—such as missing dataset category values—have become necessary. One of these methods focuses on classifying datasets according to the tags that are applied to them. The foundation of this method is a knowledge base made up of concept lattices produced for every category using the Formal Concept Analysis method. We analyze two sets of reduced concept lattices created for Ireland’s open data portal datasets in 2020 and 2021 and their usability for categorizing new datasets that were available on the portal in 2021 and 2023. Among other results, we will show that concept lattices, although reduced, can be used for a long period and still preserve the accuracy of the categorization algorithm above 90%.