Development of public health decision-making based on semantic evaluation of EHR data

Abstract

The study examines the way semantic analysis of data from electronic health records (EHRs) is performed. Implementing machine learning (ML) is challenging, EHRs are essential for acquiring medical data because they include textual information from physicians about patients' ailments and treatment plans, supporting well-informed public health decision-making. The goal of the study was to allow computers to semantically comprehend assessments, physical parts, indicators, and therapies in addition to implicitly identifying medical terms in EHRs. To efficiently identify medical phrases in electronic health records, a novel cubic support vector machine with a dipper-throated optimization (CSVM-DTO) strategy is suggested. This study uses tokenization as a pre-processing step for the data gathered from public sources. The semantic data of medical terminology is then extracted using the word-2-vector technique. Next, the named entities are found using the CSVM, and their efficiency is improved by implementing the DTO technique. Based on the experimental results, we can conclude that our suggested approach outperformed other current approaches in locating the identified entities inside the EHRs.