Method for Predicting Relations between Entities Based on Machine Learning

摘要

A method for predicting relationships between entities using a machine learning-based ranking model is disclosed. The disclosed method includes collecting data from a database; extracting a bio-object from the collected data; measuring the similarity between the extracted bio-objects through a plurality of similarity calculation methods including Word2vec, node2vec, and UMLS rules; and inputting the measured similarity value into a Learning to Rank model to infer and predict a relationship between entities. According to the disclosed method, there is an advantage of more accurately inferring and predicting the relationship between individuals.