Structured Processing Method of Medical Examination Text Reports Based on Tree Model

There are a lot of medical text reports available provided by hospital. These reports are written by doctors or generated by medical equipment specifying the clinical problems and diagnosis of patients. These reports contain very valuable information that can be useful for future patient diagnosis. But the problem occurred with these medical reports are most of them are not structured, so we will not be able to leverage full advantage of the information. This paper proposes a structured processing method based on tree model. First, data processing is performed on the text reports in order to make the data more accurate. Then establish part of speech dictionary to obtain part of speech from all words in the text reports. Semantic relationships between each sentence need to be obtained by performing dependency syntax analysis. By combining part of speech and semantic relationship we propose a rules to establish a tree model. The experiment has shown that this method can be used to achieve a good result for structuring thyroid ultrasound reports in Chinese language.