Toward Predicting Patient’s Satisfaction of Indoor Environmental Quality in Jordanian Hospitals using SVM and K-NN Machine Learning Algorithms

The indoor environment is an essential aspect of hospital design because it impacts patients’ health, well-being, and healing process. Despite that machine learning has been widely employed in various fields, few studies have looked at how machine learning may be used to improve Indoor Environmental Quality (IEQ) in hospitals. As a result, the current study uses a machine learning approach to bridge this gap. To assess the quality of the indoor environment, the researchers used mixed design methodologies. Also, self-reported data and field monitoring of environmental indicators within patients’ rooms were used to collect data from King Abdullah University Hospital (KAUH) as a sample of all Jordanian hospitals. The experiments were carried out with the same dataset for each training and testing after it was split, and the results were evaluated using the same classification metrics. It was revealed that when C equaled 0.01, the Support Vector Machine (SVM) with the linear kernel had higher accuracy in predicting patient satisfaction than other SVM kernels and the K-Nearest Neighbor (K-NN) algorithm.

I. Introduction

Hospitals are designed to provide care for patients. As a result of that, an essential aspect of designing hospitals is the indoor environment, which plays a vital role in the occupants and healthy environment [1]. Machine learning can help to provide recommendations for hospitals based on the personal data of patients [2]. Using machine learning algorithms when dealing with big data in healthcare has several advantages. One is that it can deal with different data types like demographic and non-parametric data. Also, it is an effective tool for analyzing data that have complex relationships in various fields like medicine, psychology, or others [3], [4].