Medical triage robots leverage natural language processing algorithms to provide accurate medical information and triage services, ultimately alleviating the strain on healthcare specialists. However, their effectiveness often hinges on the quality of topic assignment. This study proposes the Knowledge-Constrained Labeled Latent Dirichlet Allocation (KC-LLDA) method, which incorporates domain-specific knowledge constraints with LDA. KC-LLDA was compared with other existing similar topic extraction methods, which demonstrated that the proposed method is more suitable for topic modeling in the context of medical texts. In addition, this paper sets forth a novel hybrid method that combines supervised and unsupervised learning, leveraging the synergies between KC-LLDA and the BERT model, which results in a better learning of contextual information contained in medical texts, leading to the improvement of the classification accuracy. The obtained results highlight the fact that1 utilizing topic assignment can increase the efficiency of medical triage robots, ultimately improving the healthcare services provided to patients.
Topic assignment, Medical triage robot, Question answering system, Triage, Domain knowledge.
Jiayi FENG, Runtong ZHANG, Donghua CHEN, Lei SHI, Chenghao XIAO, "Label-based Topic Modeling to Enhance Medical Triage for Medical Triage Robots", Studies in Informatics and Control, ISSN 1220-1766, vol. 32(4), pp. 37-48, 2023. https://doi.org/10.24846/v32i4y202304
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).