A Mobile Assistant for Turkish

In this paper we present a design and an implementation of a mobile assistant application that understands and meets users' requests in Turkish language. The application is able to understand requests for a set of phone operations such as calling a contact or sending an email and requests for a set of information services such as map, weather, and traffic. The understanding of user queries relies on existing research on natural language processing for Turkish and a hybrid approach of rule-based and statistical classification methods. Our performance tests revealed high accuracy results for the operations that our application supports.