Research On Personalized Recommendation Methods Based On Domain Knowledge Maps

With the rapid development of the Internet technology,documents of the world wide web has been transformed into semantic web,semantic web has been applicated in various fields of all walks of life,and the knowledge map of the semantic web is the most intuitive and effective representation,which also makes the construction of knowledge map become the hot spot of the current research.Especially for the specific domain,to realize the personalized service needs more knowledge map as a basis for the firm,therefore,no supervision or semi supervised construction of knowledge map in specific areas has become the focus of research.In this paper,we do some research work in the field of domain knowledge mapping and information recommendation.1 a novel method based on word vector and graph model is proposed to deal with the entity disambiguation in domain-specific.Using the tool“Word2Vec" to build the model of word vector with the texts in knowledge base and the texts we grabbed from lots of tourism website.Combined with the graph of manual annotation,we used a random walk algorithm based on the graph to compute similarity auxiliarily,in order to make it accurately calculate the similarity between the words of tourism domain each other.At last,extracting keywords of the background text which the name mention is in and text which described the candidate entities in knowledge base,using the trained Word2Vec model and graphical model to calculate the similarity between the key words of name mention and the key words of candidate entities each other,choose the candidate entities which has a maximum average similarity as the target entity.Experimental Results show that it can be more accurate to achieve the entity disambiguation of domain specific.2 Personalized information recommendation in tourist area based on attribute map clustering.Firstly,based on the construction of domain entity attribute graph,then graph tier analysis domain entities are divided into different categories based on tourism,and then corresponding other conventional properties of this type of travel with the entity as the entity attribute,attribute graph clustering domain entities,at the end of this recommendation clustering the domain model of knowledge map based on experimental analysis.3 Prototype system for personalized information recommendation in Tourism.The field of information recommendation algorithm with program implementation,the keyword search crawls through the user as the input of the system will eventually meet the user interests of tourist attractions show to the user,realize the knowledge in the field of personalized recommendation.