Sucipto, Didik Dwi Prasetya, Triyanna Widiyaningtyas
Question documents organized based on Bloom's taxonomy have different characteristics than typical text documents. Bloom's taxonomy is a framework that classifies learning objectives into six cognitive domains, each having distinct characteristics. In the cognitive domain, different keywords and levels are used to classify questions. Using existing category-based term weighting methods is less relevant because it is only based on word types and not on the main characteristics of Bloom's taxonomy. This study aimed to develop a more relevant term weighting method for Bloom's taxonomy by considering the term density in each category and the specific keywords in each domain. The proposed method, called Hybrid Inverse Bloom Space Density Frequency, is designed to capture the unique characteristics of Bloom's taxonomy. Experimental results show that the proposed method can be applied to all question datasets, considering term density in each category and keywords in each cognitive domain. Furthermore, the accuracy of the proposed method was superior on all datasets using machine learning model evaluation. © by the authors.
Department of Electrical Engineering and Informatics, Universitas Negeri Malang, Indonesia; Department Information System, Universitas Nusantara PGRI Kediri, Indonesia
