DNN-Based Hospital Service Satisfaction Using GCNNs Learning

Abstract

Hospital patient service satisfaction prediction technology based on supervised learning can obtain higher prediction accuracy by training image data. This paper proposes a deep neural network based on continuous sample data to predict patient service satisfaction in hospitals, which overcomes the problem of high cost of collecting additional labeled data while improving the prediction accuracy Furthermore, our model converts labeled samples into images reflecting the characteristics of the data and trains a deep neural network on the resulting dataset, avoiding overfitting. Results based on experimental data show that the prediction accuracy of deep neural networks based on general-domain methods is higher than that of classical supervised learning methods using the same labeled data collection. At the level of practical application, the system we propose provides patients with indicators to evaluate their experience in choosing a hospital and provides a reference for hospitals to rationally arrange medical triage.

Keywords

• Deep neural network (DNN)
• hospital service satisfaction
• supervised learning
• multi-category issues
• support vector machines (SVM)