Effective Classification and Categorization for Categorical Sets: Distance Similarity Measures

Measuring the similarity between objects is considered one of the main hot topics nowadays and the main core requirement for several data mining and knowledge discovery task. For better performance most organizations are in need on semantic similarity and similarity measures. This article presents different distance metrics used for measuring the similarity between qualitative data within a text. The case study represents a qualitative data of Faculty of medicine Cairo University for theses. The dataset is about 5,000 thesis document with 35 departments and about 16,000 keyword. As a result, we are able to better discover the commonalities between theses data and hence, improve the accuracy of the similarity estimation which in return improves the scientific research sector. The experimental results show that Kulczynksi distance yields better with a 92.51 % without normalization that correlate more closely with human assessments compared to other distance measures.