A speech intelligence conversation bot for interactive media information
Jefri Yushendri, Alvian Rahman Hanif, Anneke Annassia Putri Siswadi, Purnawarman Musa, Tubagus Maulana Kusuma, Eri Prasetyo Wibowo · 2017
The customer service is needed in not only companies but also in a university. The university usually have the customer service by phone number, such as Gunadarma University that gives the Media Information Centre's phone number to do the customer service task. The growth technologies encourage the customer service to have the improvement. This research is developing the speech intelligence conversation bot as an interactive media information, a customer service improvement. This speech intelligence conversation bot is not only having the ability to listen the user's question but also has the intelligence to understand the user's question and gives the right information related to that question. The ability to listen and speak with the user is implemented using speech recognition by Sphinx-4 while the ability to understand and find the related information is implemented by using the chatbot system with the modificated MegaHal style for developing it. The accuracy each ability of this speech intelligence conversation bot are intelligence conversation bot are 80,0% for speech recognition, 89,3% for chatbot system, and 100% for response generation. The average accuracy fot the whole system is 89.7% However, it needs more improvement in both speech recognition ability and chatbot system ability, especially for speech recognition since it has the lowest accuracy number among the other abilities.