Research And Application Of Construction Engineering Information People Relationship Recommendation System Based On Graph Database

With the quick advancement of information technology,users are presented with an abundance of information on the Internet,making it challenging for them to quickly and effectively locate the information they require,which creates the issue of information overload.Recommendation systems were created because conventional solutions like classification and search engines are useless unless the user has a clear need for them.In this study,the idea of a recommendation algorithm is applied to a fresh viewpoint on people in the field of construction engineering information,which recommends the relationship links of target people to users by building multi-dimensional relationships between projects,businesses,contacts,and users,and helps users solve the problem of how to get accurate recommendation links of people when there is too much project information and users cannot browse all the information.This paper firstly provides a comprehensive discussion of the construction engineering information people recommendation system,and through the analysis of the functional requirements of the system and the elaboration of the current domestic and international research results and mainstream algorithms,it studies and compares different recommendation algorithm strategies with factors such as user content and network topology,and finds out the suitable people recommendation strategy for the construction engineering information industry based on the research results.In order to improve the accuracy of people link recommendation,the information attributes and labels carried by the target contacts’ related companies,related projects,intermediaries they may know,and intermediate relationships are clustered and initially ranked,and then Graph DB graph database is used to try to find the association relationship in a large amount of data and to realize the extension function by quickly extending the relationship.Based on the node similarity calculation in the people recommendation algorithm,combined with the link prediction methods of common neighbor,path and classification,this paper finally implements the construction engineering information people recommendation system.Through comparison tests,the experimental results show that the method proposed in the paper can effectively improve the accuracy rate and interface response speed of the recommendation system.The front-end framework of this paper adopts vue.js,the back-end framework adopts Ruby on Rails,and the database adopts Postgre SQL.In contrast to other systems in the same field,this paper introduces statistics based on user-initiated marking behaviors in the people recommendation strategy,as well as detailed supplementary descriptions of recommendation information.By providing users with clear descriptions of specific links to direct and indirect people,users can quickly verify the accuracy of recommended people relationship data and increase the accessibility of information.After the improvement,the user login activity of this system showed a significant increase,and the performance load of the recommendation module also improved significantly.