Implementation of Logistic Regression and Naïve Bayes Methods in Analyzing Employee Sentiment at Dr. Iskak Hospital Regarding Organizational Performance

Abstract

This study combines the McKinsey 7S model and machine learning-based classification techniques to study employee sentiments towards organization performance at Dr. Iskak Hospital, Tulungagung, East Java. The data were obtained by interviewing 56 employees and managers which can be classified into positive, neutral, and negative class by referring to 7S aspects in McKinsey model as previously defined. Term Frequency-Inverse Document Frequency (TF-IDF) is used for feature extraction in which it is implemented for translating words into numeric vectors. There are two classification methods namely Logistic Regression (LR) and Naive Bayes (NB) applied and compared under two simulations with variations in training-test dataset split (70:30 and 80:20) by applying precision, recall, F1-score and accuracy as evaluation measures. It can be concluded that both methods produced the same evaluation scores in the two simulations. In the first simulation, both methods yielded an accuracy rate of 71 percent, which increased to 75 percent in the second simulation. The results obtained from the implementation of these two methods provide a broader perspective on organizational climate, which will serve as valuable insights for data-driven decision-making and corrective actions within a healthcare organization.