Latent Dirichlet Allocation for Medical Records Topic Modeling: Systematic Literature Review

Abstract

The fast growth of Electronic Medical Records (EMR) has improved its functionalities and increase its use in secondary functions. EMRs can be used to improve the quality and capacity of physicians and medical students. It is done by using EMRs as a data source for researches and learning. A lot of studies have been done in this area. Various methods are also proposed. One of the methods used in medical records retrieval is Latent Dirichlet Allocation (LDA). Thus, research on LDA for medical records has also been carried out quite a lot. Unfortunately, those researches are still scattered and make it difficult to know the state of the arts of utilizing LDA in the medical records field. Our research intends to explore the studies of the LDA method progress in the medical field. To better present responsible literature study, our research used Systematic Literature Review as a research methodology. The literature review was conducted on studies from 2015 to 2021. The result of the literature review showed that the LDA method had been used in several research topics. And most of the literature used private datasets in their experiments. Some researchers also added a modification in the LDA method. Even though the LDA method proved a great method in medical fields, it has several limitations that need to be overcome.