Graph data analysis searches through patterns of data and indicates the most optimal path
or result that best suits the users’ requirements.
Graph Indicator®
CASE STUDIES
Knowledge Graph in Manpower Matching System
In order to manage ever-increasing data effectively and efficiently, the Center for International Development (CID) of the Korea Development Institute (KDI) established an advanced knowledge management system with Bitnine’s AgensGraph. Reinforced with an optimal graph query language, visualization tool, and graph algorithms, the knowledge graph elevated the value of data and evolved the knowledge management system to another level. The newly built manpower matching system intuitively connected relationships between various data accumulated during overseas projects that KDI experts have participated in.
AI-based Personalized Learning Recommendation
Company E used AgensGraph to provide fully personalized education services for young students. This service took into account each student’s academic achievement, behavioral patterns, and personal preferences, and then suggested an effective way to study. Unlike simply determining students’ progress based on their test scores, this service provided a conceptual framework of the school subjects in a network form and recommended the most effective course for each individual student.
Graph-based Recommendation System
Collaborating filtering and content-based filtering are widely known recommendation algorithms. However, each algorithm has its own limitations which becomes a challenge for the recommendation engine. A graph DB is proposed as a new solution to the recommendation system. Coupled with graph algorithms and various recommendation techniques, data of user’s content can be evaluated within the graph structure.
Recommending Researchers & Thesis based on Graph Analysis
With graph DB, the research institute manages a researcher recommendation system. Research activities are analyzed, while also removing any repetitive data on the researcher and thesis paper, and distinguishing researchers with the same name.
