Signs and Symptoms Tagging for Thai Chief Complaints Based on ICD-10

Abstract

This paper presents a natural language processing (NLP) approach to construct signs and symptoms corpus in order to identify signs and symptoms recoded in a Thai chief complains (CCs) based on the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) form. We define our native language "Thai language" as the natural language in our works thus the challenge is how to apply NLP concept that is originally designed for English language. We start from tokenization to extract Thai token from Thai chief complains, and then the tokens is analyzed in order to assigning a specific tag in terms of ICD-10 code.