An automatic health surveillance chart interpretation system based on Indonesian language

Abstract

The Indonesian communities have the difficulties for interpreting the chart image, especially the health surveillance chart image. This condition is impacted on late awareness of disease threat. In Indonesia, the chart interpreter system based on Indonesian language is available, but it still has the limitation in generating the textual representation (i.e the health surveillance summaries) in the case of the user involvement. This paper proposed the automatic chart interpreter system as the Natural Language Generation (NLG) system adopting Data-to-Text architecture for overcoming the intended problem. In other words, it emphasizes on the automation system, where the automation system refers to the ability of system for determining the related events and the inference of the related events as well as to generating the related summaries without requiring the manual input from the users. Based on the experiment result, the proposed system (HS-CISys) can perform the automation with involving Data Interpretation and Micro Planning. Therefore, Data Interpretation and Micro Planning ensure that the proposed system can generate the health surveillance summaries fully automatic without relying on the manual input from the users.