Data-driven Construction of Rural Persona Based on Text Mining and Knowledge Association

The implementation of rural revitalization strategy in China makes “rural tourism” become one of the key industries of rural development. A comprehensive mining of the characteristics of various villages is helpful to promote the overall development of rural economy and realize the modernization of agriculture and rural areas. Based on big data and relevant policy texts, this paper develops a labelling framework, and proposes a deep mining of the rural themes and labels by integrating the deep learning-based text classification and LDA topic mining. In the study, a knowledge graphs of 4,814 entities and 7,369 relationships including villages, scenic spots and featured products is constructed using multi-source heterogeneous data, and we also realizes label fusion and system construction. Through the visualization and analysis of the rural panoramic feature persona, the development modes and the features of villages are described in different aspects. By the construction of rural personas, the study presents the typical features of key villages in rural tourism in China from multiple perspectives, which provides a strong data support for building the recommendation system of online tourism websites and is of great value for promoting the rural tourism.