Personalization Recommendation Model Research Of Agricultural Information Service Based On Ontology

Agriculture is one of the pillar sectors of China's social and economic development.With the database as the core,agricultural comprehensive information service platforms currently provides convenient,efficient,comprehensive interactive,professional,personalized agricultural information services for enterprises and people by means of highly integrated agricultural practical technology,market supply and demand,production and marketing docking,production and operation status,and other information resources,and by taking the Internet and all kinds of terminal equipment for communication and sharing channels.Currently,there are still some deficiencies in the construction and operation of most agricultural comprehensive information service platforms:First,the matching degree between huge agricultural information and agriculture-related user demand in the platform is not high,which can be attributed to failure to accurately delineate user intentions and implement user personalized recommendation information;second,agricultural integrated information service platform lacks knowledge integration and relevance,and the information integration is poor.Entities can be more closely linked to entities by the construction of ontology-based agricultural domain knowledge graph;thirdly,user engagement on the platform is poor,where system cannot accurately analyze the user or accurately recommend the user's concerns.In view of the above problem one,read user intentions by building user portrait models,specifically:Establish a user appeal model,and make use of machine learning technique to develop the demand label caliber for the related problems concerned by users.A TextCNN algorithm based on word-weighted representation is proposed.This model can predict users' intention according to the content of their posts,which is usually accompanied by emotional expression.In this regard,the method based on emotion dictionary is used to analyze the emotions of users in the posts on the basis of constructing the persona of users' demands,and describe users' emotions and demands together.Also,it is conducive to better understanding users;On the other hand,the relevant terms that users are interested in can be obtained to describe users' online behavior through the web pages and search terms that users care a hang,thus enriching users' personas.In allusion to the second problem,there are too many unstructured texts without complete information closed loop.Accordingly,based upon agricultural ontology knowledge,the entity information based on agriculture is constructed,and the agriculture-related information is formed in graph through entity.The specific practices are:The agricultural encyclopedia data were obtained according to agricultural policy information and agricultural related data information to use BERT+BiLSTM+CRF to identify entity.The agricultural information knowledge graph based on ontology was constructed by combining the identified entities with the original text content,and the agricultural information knowledge was organized into knowledge graph.In allusion to the third problem,currently most agricultural information individuation service is kind of stiff,in which everyone has their own view independently.Hence,a personalized recommendation algorithm based knowledge graph was proposed by pushing it to users in the form of "user+information" combined with the construction of users' personas in the first problem and agricultural information knowledge graph in the second problem.In this way,the information pushed by the system is the information that users are concerned about and potentially interested in,thereby enhancing user engagement and service improvement.In conclusion,this article focuses on the personalized recommendation model of agricultural information service based on ontology concept and concretely studies the entity recognition of user portrait and agricultural information as well as user's retrieval of agriculture-related information and the implementation of recommendation algorithm in the process of personalized recommendation.The above study has provided technical support for personalized recommendation of agricultural information service.