Prostate cancer management with lifestyle intervention: From knowledge graph to Chatbot

BackgroundPersonal lifestyle is an important cause of prostate cancer (PCa), hence establishing a corresponding knowledge graph (KG) and a chatbot is a convenient way for preventing and assessing risks. The chatbot based on a KG of PCa-associated lifestyles will be helpful to PCa management, then save health care resources in the ageing society. ResultsBased on our established knowledge base, we define entities and corresponding relationships to construct the PCa-associated lifestyles KG for visualization by importing the triples into the Neo4j graph server. The dialogue system uses the Flask framework to determine the classification of questions through entity recognition and relationship extraction and later uses the query template to search the answers from the PCa-associated lifestyles KG. The PCa-associated lifestyles KG contains 11 types of entities and 14 types of relationships, the total number of nodes and links is 21 546 and 66 493, respectively. Also, the entity "Lifestyle", "Paper", "Baseline" and "Outcome" contain multiple attributes. The established chatbot can answer 12 types of basic questions and predict the probability of a certain lifestyle resulting in a certain PCa. The chatbot is available at . ConclusionA chatbot based on PCa-associated lifestyles KG was constructed to help researchers, physicians or patients learn more about PCa lifestyle management interactively.