Automatic system for Thai query distribution

Abstract

In order to develop an automatic system for Thai query distribution, sometimes keyword spelling in the query is matched with a name of an unrelated department, which causes wrong distribution. This thesis proposes an automatic Thai query distribution system by considering context relation between the query sentences and the department context. More specifically, the context relation takes the keyword frequency and the keyword priority into account so that the precise relation between the query sentences and the departments can be identified. The system firstly inputs a query sentence in a natural language phrase. Then, the system computes the context relation and matches the right department. The experiment with five divisions of Mae Fah Luang University and 115 queries shows that the proposed system provides 86.09% distribution accuracy.