Conversational Recommender System Based on Functional Requirements and Technical Specifications

The online shopping is very inviting for people to save travel time and attained cost. When a person seeks to buy something, the easier way to find the product is by asking directly to the seller. However, when the person tries to buy things online, the communication with the seller can be limited and not effective for the seller. Among these people who try to buy product, some of them might be familiar with the specification of the product they want, but some people also might not be familiarwith the specification of the product. Our previous study has compared two models, the first one used functional requirement and the second one used technical specification. That study suggests combining the functional requirement andTechnical Specification into one model. To address this problem, this research proposes a conversational recommender system based on functional requirements and technical specifications using knowledge- based method. We use ontology to represent the system knowledge. To make the chatbot able to understand the user's message, the system is implemented on chatbot by using DialogFlow as the natural language processing (NLP) tool. To evaluate the system, we use user satisfaction evaluation method to reach the success rate of the system based on user's subjective opinion. Based on the result of the evaluation, it is concluded that theusers are satisfied with the chatbot with score above 83 % on 7/8 statements.