The rising multifaceted nature in healthcare, exacerbated by a maturing populace, brings about ineffectual basic leadership prompting inconvenient impacts on care quality and raises care costs. Therefore, there is a requirement for brilliant choice emotionally supportive networks that can enable clinician's to settle on better educated consideration choices. Choices, which are not just founded on general clinical information and individual experience, yet in addition lay on customized and exact bits of knowledge about future patient results. A promising methodology is to use the progressing digitization of healthcare that creates phenomenal measures of clinical information put away in Electronic Health Records (EHRs) and couple it with present day Machine Learning (ML) toolset for clinical choice help, and at the same time, grow the proof base of medication. As promising as it sounds, acclimatizing total clinical information that gives a rich viewpoint of the patient's health state accompanies a large number of information science challenges that hinder productive learning of ML models. In this paper we use Glove and NB models for classification of EHR data. The proposed model works with two different kinds of medical text notes one is EHR data covers the different types of ICU admissions. And make classification of ICU data with respect to mortality and re-admission of emergency departments. And second thing is working with the emergency cases of heart attack patients who are readmitted and who are passed away in ICU. The created models are observationally assessed with regards to anticipating unfavorable clinical results, for example, mortality or medical clinic readmissions.
Volume 12 | Issue 3
Pages: 295-303
DOI: 10.5373/JARDCS/V12I3/20201194