The traditional evaluation of public transport services is mostly limited to some physical indicators of facilities, people's views on public transport services are rarely taken into consideration. There are a large number of public transport service reviews on the Internet. Natural language processing technology can easily provide insight into people's opinions for public transport, thus forming some valuable social indicators. Knowledge is formed directly through entity recognition, and it is further divided into five dimensions after artificial division, together form the social indicators. Combine the physical indicators, we structured two kinds of indicators to form specialized evaluation knowledge, and constructed the public transport service evaluation knowledge graph. It is convenient to evaluate the service quality, so as to more comprehensively reflect the problems of public transport system. On the basis of improving public satisfaction, it forms a service closed loop which plays a positive role among public transport operating companies.

CICTP