tailieunhanh - Báo cáo khoa học: "A Hierarchical Approach to Encoding Medical Concepts for Clinical Notes"

This paper proposes a hierarchical text categorization (TC) approach to encoding free-text clinical notes with ICD-9-CM codes. Preliminary experimental result on the 2007 Computational Medicine Challenge data shows a hierarchical TC system has achieved a microaveraged F1 value of , which is comparable to the performance of state-of-the-art flat classification systems. | A Hierarchical Approach to Encoding Medical Concepts for Clinical Notes Yitao Zhang School of Information Technologies The University of Sydney NSW 2006 Australia yitao@ Abstract This paper proposes a hierarchical text categorization TC approach to encoding free-text clinical notes with ICD-9-CM codes. Preliminary experimental result on the 2007 Computational Medicine Challenge data shows a hierarchical TC system has achieved a microaveraged Fl value of which is comparable to the performance of state-of-the-art flat classification systems. 1 Introduction The task of assigning meaningful categories to free text has attracted researchers in the Natural Language Processing NLP and Information Retrieval IR field for more than 10 years. However it has only recently emerged as a hot topic in the clinical domain where categories to be assigned are organized in taxonomies which cover common medical concepts and link them together in hierarchies. This paper evaluates the effectiveness of adopting a hierarchical text categorization approach to the 2007 Computational Medicine Challenge which aims to assign appropriate ICD-9-CM codes to free text radiology reports. Pestian et al. 2007 The ICD-9-CM 1 which stands for International Classification of Diseases 9th Revision Clinical Modification is an international standard which is used for classifying common medical concepts such as diseases symptoms and signs by hospitals insurance companies and other health organizations. The 2007 Computational Medicine Challenge was set in 1 see http nchs a billing scenario in which hospitals claim reimbursement from health insurance companies based on the ICD-9-CM codes assigned to each patient case. The competition has successfully attracted 44 submissions with a mean micro-averaged Fl performance of . Pestian et al. 2007 To the best of our knowledge the systems reported were all adopting a flat classification approach in which a dedicated .