Prediction of Hospital Admission Using Machine Learning

Abstract

The process of seeking hospital admission is fraught with a
variety of difficulties that patients must navigate. If a hospital is
overcrowded, individuals might expect to wait in line fo r many
hours just to be admitted to the facility. It is not suitable t o use
this method in the Emergency Department. Patient will be
admitted to the hospital's emergency department if his or her
condition is considered to be serious. As a result of the
increasing demand, we'll need to develop new techniques to
optimise patient flow and alleviate congestion at the hospital.
Thus, data mining algorithms will find a humorous technique to
anticipating emergency department admissions as a
consequence. The prediction algorithms that we looked at were
the Naive Bayes, Random Forests, and the Support Vector
Machine, among others. We'll need to identify a few parameters
that are related with hospitalisation, such as age, gender,
systolic and diastolic blood pressure, diabetes, prior d ata from
the preceding month or year, and the date of admission, in order
to make the forecast. Aside from that, we go into great detail on
the methods that we used to accomplish our goals. By
categorising data into categories, we may enhance prediction
accuracy by boosting prediction accuracy. We use the Rand om
Forests approach to do this. The Naive Bayes approach is used
to estimate the likelihood of each attribute occurring, which
assists in the prediction of the outcome of the experiment. Using
the Support Vector Machine, you may classify your input d ata
into one of several categories, which will help you forecast your
output data.