Semantic Similarity Measures for Knowledge Engineering: Experiments on Umls, Wordnet and Biomedical Corpus

Knowledge management has been considered in the past decades applying techniques to cope with organizing data, information and knowledge. It implements methods to manage knowledge on papers as well as digital ones. Meanwhile, birth of the World Wide Web despite all of its advantages has initiated a number of issues for the researchers due to massiveness of the data on the Web. Therefore, the new knowledge engineering techniques must be automated to save time and effort of man power with considering the Web with a shared understanding of the data among all of its components. To achieve accurate and integrated definition of all available data, machines need to make a unique understanding of all discrete data sources. This book is aimed at presenting existing Measures of Semantic Similarity for resolving foregoing issue. These measures are also useful in tasks such as text categorizing, machine translation and information retrieval. Furthermore, this book introduces two new normalized functions for measuring semantic similarity between two concepts based on first and second order context and information content vectors computed from MEDLINE as the biomedical corpus, UMLS and WordNet.