On improving open dataset categorization

Under influence of open government and data transparency initiatives, a variety of institutions have published a significant number of datasets. In most cases, open data portals are used as tools for publishing purposes. Further, open data portals have the ability to make data more accessible by categorizing datasets based on different criteria like publishers, institutions, formats, and descriptions. For these purposes, portals take advantage of metadata accompanying datasets. However, a part of metadata is often missing which makes it harder for users to find appropriate datasets and obtain the desired information. As the number of available datasets grows, this problem becomes easy to notice. This paper is focused on first step towards decreasing this problem by implementing components capable of suggesting the best match for the category where an uncategorized dataset should belong to. Our approach relies on dataset descriptions provided by users within dataset tags and uses formal concept analysis to reveal shared conceptualization originating from tags' usage. Since tags represent free text metadata entered by users, in this paper we will present a method of optimizing their usage through means of semantic similarity measures based on natural language processing mechanisms. Finally, we will demonstrate the advantage of our proposal by comparing concept lattices generated using formal concept analysis before and after optimization process and use generated structure as a knowledge base to categorize uncategorized open datasets.