Patient-centered doctor recommender system of online health communities: A multidimensional sequence alignment approach

Abstract

This study develops a patient-centered Doctor Recommender System (DRS), facilitating patients' doctor choice performance by calculating doctors' "biological distance" based on Multidimensional Sequence Alignment (MSA) method. The distance is measured by doctor-patient interaction attributes, which are acquired from a combination of Levenshtein Distance and opinion mining technology. The XGBoost is used to identify the importance of these attributes. The precision of patient-centered DRS is 72.60 %, and the recommended doctors are significantly different from the doctors chosen by patients lacking medical knowledge. It not only furnishes practical guidance for OHCs in increasing patients' consulting satisfaction, but also illuminates the important properties in improving doctors' popularity.