<![CDATA[Study of atmospheric are currently growing rapidly to analyze the negative effects of climate change, weather and air quality. Unstable atmospheric conditions have a negative impact, as extreme weather. The combination of technology and analysis of atmospheric conditions is currently developing rapidly. While atmospheric research using machine learning technology and algorithms is advancing swiftly, challenges persist in identifying the optimal machine learning model for precise cloud type classification. The application of Machine Learning algorithms in atmospheric research has been widely carried out to predict wind direction and cloud imagery to detect weather using satellite data. Detecting cloud type is important for predicting the upcoming weather. However, to detect the type of cloud, it is necessary to choose the algorithm with the best performance. This study applies the Convolutional Neural Network (CNN) with EfficienNetB3 method, Support Vector Classifier (SVC), XGBoost Classifier (XGB), Extra Tree Classifier (ETC), Random Forest (RF), and Decision Tree (DT) algorithms in classifying cloud images to detect clouds type. The method used in this research involves an experimental approach in the hope of gaining a deeper understanding of the factors that influence the performance of machine learning models in classifying cloud types. The dataset used in this research is 1500 cloud data (1200 training data, 300 testing data). Researchers conducted a comparison of algorithms to find out the best algorithm performance in classifying cloud type images. The results showed that doing the CNN algorithm showed better performance with an average accuracy got of 81.03% compared to the SVC algorithm (34.44%), XGB (33.79%), ETC (39.25%), RF (36.18), and DT (29.35%). Our contribution to this research is that we compare machine learning algorithms to detect cloud types along with the impact and characteristics of cloud types from the prediction results.]]>
