Research On The Method Of Information Recommendation In Chinese Tourism Field Based On Entity

In this paper, the tourism personalized recommendations were studied, unstructured free text field were specific instances of domain concepts, entity identification and extraction between attributes and attribute values, build Yunnan tourism ontology knowledge base for tourism in the tourism user providing personalized information recommendation forums. Examples of the field, between the attributes and attribute values to identify and extract entities, to build domain-specific ontologies and recommend follow-up information has practical significance. In this paper, the work done in the following areas.1. Markov logic network concept in the field of tourism instance, between attributes and attribute values to identify several types of entities, including the corpus preprocessing, feature selection, format conversion and other processes. First-order logic formulas to represent the range features, characteristics and correlation length near the associated far-length features, and features integrated into the three Markov logic network concept instances, attributes and attribute values named entity recognition in the field of tourism. From the experimental results, the concept of instances, attributes and attribute values section feature integration, nearly the length of the correlation length associated with characteristic features and far Markov logic network in the Chinese tourism free text field named entity recognition is feasible, and has good extraction performance.2. Knowledge introduce a conceptual model composed of five elements and build the knowledge base of the five criteria Grunder raised. Using ontology tool protege and a brief introduction, last, tourism body of knowledge of the design process, culminating in a tour ontologies. Detailed design of the body of knowledge in the concept of tourism instance, kilometers classes, properties, property values, constraints and class structure.3. Facing serious information overload for tourism user problem, collaborative filtering recommendation method based posts of tourist information. After analyzing the characteristics of the knowledge base of information recommended property, the first use of a blend of user reviews PageRank algorithm to judge the behavior of the importance of the individual user, the main consideration of the time response relationship between the response among users as well as individual users. Then the PageRank score as high user clustering K-means clustering centers, users and user recommendation system then obtained by clustering collaborative filtering algorithm to calculate the similarity, combined with the user’s PageRank score, select the user relevance information as a result of the higher recommended.