Combination of Inner Approach and Context-Based Approach for Extracting Feature of Medical Record Data

Abstract

Changing from a legacy Health Information System (HIS) to a modern HIS creates a problem of migration. Particularly, it requires us to handle unstructured data. In this paper, we proposed a new approach which is used to detect keywords from textual documents, and it involves two stages. Firstly, we study extracting features from the words by exploiting the relation between their characters. The following stage is presenting a combination of inner approach and context-based approach in order to make this extraction. The method is tested with MIMIC-II dataset and, in our problem, it shows a better result compared to old methods. We believe that it can be applied into natural language processing problems in other fields.