Hierarchical Document Classification by Conceptual and Semantic Similarities: CSS-HDC

The exploitation of syntactic components and the semantic environment has long been a significant issue in the context of data mining and information retrieval, particularly with regard to text data. This issue's effectiveness has been noticeable in completely unique tasks, such as supervised learning of text data. So far, extra syntactic or semantic information has been used only distinctively. With inspiration from our previous work, which successfully defined concept labels for supervised learning, we propose a hierarchical document categorization based on conceptual and semantic relevance in this paper. The conceptual relevance is validated by the concept labelling approach developed in our previous research article. The main contribution of this paper is to investigate semantic relevance by estimating the correlation between concept categories based on activity labelling. The findings of the empirical study concluded that the developed model is promising for significant classification of given documents based on conceptual semantic relevance.