<![CDATA[Indoor carbon dioxide (CO2) concentrations can negatively impact human health, making it essential to predict their hazardous levels. This study applies the Adaline Neural Network, an artificial neural network algorithm, to predict the classification of CO2 as dangerous or non-hazardous indoors. The main focus of this study is to explore the use of early stopping parameters with patience values to minimize the number of iterations required in the training process. The data used in this study consisted of 596 training data, 85 validation data, and 170 test data. Based on the experimental results, Adaline Neural Network achieved an accuracy of 97.06% on the test data using a learning rate of 0.02. In addition, observations of the early stopping parameters showed that there was no significant change in the validation Mean Squared Error (MSE) value and the test data classification results even though using 5 and 10 patience and 500 iterations. These findings indicate that the use of early stopping can speed up the algorithm's stopping process at the eighth iteration without reducing prediction performance.]]>
Copyrights © 2026
