Detection of Interaction-based Knowledge for Reclassification of Service Robots: Big Data Analytics Perspective

Abstract

With the advancement of artificial intelligence technology, the robot industry in human- robot interactive service has rapidly developed. The purpose of this paper is to uncover user acceptance of human-robot interactive service robots based on online reviews. Extract reviews the public service robots and the domestic service robots from YouTube uses word2vec, sentiment classification, and LDA (Latent Dirichlet Allocation) analysis methods for research. The results show that in the interactive technology, the public service robots, the domestic service robots, and the service robots can well receive the user’s speech, gestures, and understanding of emotional states and navigating with and around. However, collaborating with humans, users may be more fearful and worried. At the same time, the positive topic of the public service robots is experience value, and the negative topic is system quality. The positive topic of the domestic service robots is anthropomorphism, and the negative topic is perceived intelligence.