
Increased Accuracy Of Sequence To Sequence Models With The CNN Algorithm For Multi Response Ranking On Travel Service Conversations Based On Chat History
Author(s) -
Wahyu Wijaya Widiyanto,
Uli Rizki,
Edy Susena
Publication year - 2020
Publication title -
jurnal infotel/jurnal infotel
Language(s) - English
Resource type - Journals
eISSN - 2460-0997
pISSN - 2085-3688
DOI - 10.20895/infotel.v12i2.470
Subject(s) - ranking (information retrieval) , sequence (biology) , computer science , consistency (knowledge bases) , chatbot , base (topology) , service (business) , information retrieval , construct (python library) , algorithm , artificial intelligence , knowledge base , learning to rank , machine learning , data mining , mathematics , mathematical analysis , economy , biology , economics , genetics , programming language
Building a chatbot cannot be separated from the knowledge base. The knowledge base can be obtained from data that has been labeled by the developer, documents that have been converted into pre-processing data, or information taken from social media. In this case, the data used as knowledge is chat history. In the chat history there are certainly many variations of answers and allowing a question to give rise to many answers. To overcome these multi responses, response was made. The existence of ranking, of course the response desired by the user will be more appropriate. Challenge in ranking is how to get the essence a question and find pairs questions and answers from the data. This can be solved by a sequence to sequence model. However, the problem that will arise is the consistency of the answers. The existence of a lot of chat history certainly raises many explanations, even though the question's essence is the same. For this reason the CNN algorithm as a solution to the problem. This research uses convolutional sequence to sequence which will be applied for ranking responses. We compare the efficiency of this model. By making comparisons, this model shows an accuracy of 86.7%