Abstract
During the process of designing and implementing a working environment, there is a need to guarantee adequate conditions for future workers' health and well-being. This article addresses the classification of employees characterized by several basic input variables (gender, age, class of work). The investigated variable was the health of employees. This article aims to create a prediction classification model using the classification tree, which can be used to classify new cases into appropriate classes as accurately as possible. Objective measurements of microclimatic parameters were performed by the Testo 435 instrument. The subjective evaluation was performed by a questionnaire survey formed from the training group of 80 respondents and independently verified by the test group of 80 more respondents. The result confusion matrix shows that the number of correctly classified respondents was 69 from a total of 80 respondents. The overall accuracy was A C = 0.863 , which means that the likelihood that respondents are properly classified in the correct health class is 86.3%. Based on the model obtained using the classification tree, we can classify respondents into the relevant class for their state of health. The respondent is classified into the class of work for which particular health and working conditions are most likely.