Research On Classification Framework Of Patient Complaints Based On Hierarchical Topic Model And Generic Integration

[Purpose] The existing classification framework for patient complaints is usually a summary of expert experience and has certain limitations.According to the specific situation of domestic patients,combined with the existing classification framework of patient complaints,this study uses the method of machine learning hLDA level topic modeling to mine the subject complaints to form a more scientific and rational classification framework,which is better for better Evaluate the quality of medical services.[Methods] Using literature analysis method,summarize and summarize the existing patient complaint classification framework;use HLDA-based method,use python and related toolkits to mine the topic,and then use topic-by-point mutual information technology to analyze topic mining.As a result,the corresponding classification framework of patient complaints was summarized;the classification framework of the two patient complaints was integrated by using the integrated method in qualitative analysis.[Results] Through the use of hLDA hierarchical topic modeling method,the topic mining,constructed a classification framework based on patient complaints,including 5 major categories,7 subcategories.Visual analysis showed that patient complaints were mainly concentrated in clinical services,environment,and management,accounting for 88.37%;Using the method of generic integration in qualitative analysis,the internationally accepted HACT classification framework and patient complaints based on hLDA-level topic modeling After the classification framework is integrated,a new classification framework for patient complaints is obtained: clinical service quality,medical error,medical environment,life support,institutional process,management,communication with patients,respect and patient rights..[Conclusions] This study uses the hLDA hierarchical topic model algorithm to mine the subject complaints of a hospital,and then summarizes the excavated topics,and obtains a two-tier patient complaint classification framework,with commonly used classification frameworks,such as HACT,there are certain differences.Then,after integrating the internationally accepted HACT classification framework and the classification framework of patient complaints based on hLDA level topic modeling,a new classification box for patient complaints is obtained,which is closer to the true distribution of the patient complaint topic and more comprehensive.It is conducive to a more scientific and reasonable assessment of the quality of medical services.