Analisis Perbandingan Metode K-Nn Dan Decision Tree Dalam Klasifikasi Kenyamanan Thermal Bangunan

Abstract

Thermal comfort and good air quality can have a positive influence on a person's health and activities. Human knowledge will determine the right quality for a building or room that can only be used at a minimum first if it is based on the value of temperature and humidity standard measuring instruments that display raw data so that a system is needed that makes it easy for humans to determine thermal building comfort. Most systems that discuss thermal comfort revolve around temperature and humidity monitoring still do not use a machine learning model to maximize data analysis. In this study, a comparison of machine learning models is designed that is able to provide a classification to predict the category or label given to a data set consisting of the thermal comfort variable. The output produced in this comparison is the result of labeling predictions from 3 temperature comfort levels by testing 30% praise data and 70% training data. From the results of the segmentation accuracy level using K-Nearest Neighbor, the accuracy reaches 100% with the highest accuracy at a value of K = 1, while the Decision tree comparison method also gets an accuracy value that reaches 100%, this shows that thermal comfort can be applied in the classification machine learning method