TEMU KEMBALI INFORMASI UNTUK PEMBANGKITAN BASIS PENGETAHUAN DARI TEKS BEBAS YANG DIGUNAKAN OLEH AGEN PERCAKAPAN BAHASA ALAMI
Conversational agent is an intelligent agent which has an ability to respond with an intelligent way to user request. In natural language dialogue interaction, conversational agent is often required. In this research we develop an IR (Information Retrieval) module to be used by a knowledge-based conversational agent.
Usually Knowledge Base (KB) is given by inputting data using structured information, such as KIF(Knowledge Interchange Format), KQML (Knowledge Query and Manipulation Language), or Semantic Web RDF(Resource Description Framework). Today with The advance of Internet, information in the form of web or free text can be found easily.
In this research, KB is generated from Indonesian free text. World tourism is used as domain knowledge. The free text is processed using an IR module employs keyword-based feature i.e. geographical name. KB utilizes semantic network to relate facts with semantic relation for example “things(isA, has)”, “spatial (locationOf)” and question words for example “what (things)”, “where (spatial)”, and “how(situation)”. Application of this research will enable user to enrich knowledge-based conversational agent for geographical domain using easy-to-write description of any tourism place in the form of Indonesian free text.
Keywords : Agent, Information Retrieval, Knowledge Base, Free text.