Abstract:At present, the popular agricultural information services are mainly telephone direct consultation, centralized technical training and expert on-site guidance in China. Due to the limitation of time and space and manpower, there is a lack of timeliness and convenience. Through the research and development of Android agricultural technology intelligent question answering robot APP, agricultural information service can be provided for farmers. Crawlers were used to collect a large number of agricultural technology Q&A data on Internet platforms, which were preprocessed to form a corpus. The CRF model was trained to recognize the agricultural technology named entity after automatically labeling the corpus features. According to word frequency and information entropy, the evaluation index of named entity was calculated to construct the triple knowledge base of “crops, pests and pesticides”. The knowledge base was imported into Neo4j to establish the agricultural technology knowledge map. The algorithm of named entity recognition and knowledge map query recommendation was integrated in Android to solve the problem of keyword recognition and query result recommendation. This question answering system can provide a intelligent solution for agricultural technology Q&A, which had a high degree of automation and application value.
